forked from enderofwings/NexusOS
AthenaandClaude Sonnet 5
663e540b6b
Adds synapse/slash_commands.py: a chat message that's nothing but /tool_name(arg=val, arg=val) dispatches straight through tools.dispatch(), skipping model selection, RAG/playbook context assembly, and the ask-policy approval round-trip entirely. A human typing this IS the approval - there's no one else to ask - so it's a deliberate, reviewed bypass of the approval step specifically, not of anything a tool validates internally (path boundaries, size caps, Curry's own sandbox checks all still run). Argument values parse via ast.literal_eval only: strings/numbers/bools/None/literal containers, no names, no calls, no attribute access - a malformed or hostile-looking argument fails to parse rather than executing anything. Wired into chat_stream_endpoint (main.py) as an early short-circuit, before any of the RAG/model-selection work that a slash-command doesn't need. Web needed no changes (it already forwards raw text unchanged); the TUI previously swallowed every leading "/" locally and never reached the backend with it, so tui_app.py's _handle_slash now falls through to _start_chat for anything shaped like a tool call while still handling its own local meta-commands (/help, /model, /new, ...) exactly as before. Also finally wires Curry in as ten real tools (curry_declare_constant, curry_get_constant/_latest, curry_list_constants, curry_retire_constant, curry_declare_function, curry_get_function, curry_list_functions, curry_call_function, curry_retire_function) - deferred from the vendoring pass. The five write/execute ones are ACTION tools in the same always-ask-regardless-of-global-policy floor as edit_source (ALWAYS_ASK_ACTION_TOOLS, generalized in tools.py from the old self_edit-only ALWAYS_ASK_TOOLS so future tool families share one place to register into). curry_call_function is gated as an action for the same reason run_snippet is: it executes code, even sandboxed. Fixed a real bug surfaced while wiring this up: curry_db is a long-lived singleton holding one sqlite3 connection (unlike NexusOS's own memory store, which opens/closes a fresh connection per call specifically to dodge this), and sqlite3 forbids using a connection from a different thread than created it. That's a non-issue in production (uvicorn's single event-loop thread), but Starlette's TestClient runs the ASGI app through an anyio portal thread, so it broke immediately under test. Fixed at the source (curry_core.py, Curry.__init__) with check_same_thread=False, documented as a second deliberate vendoring deviation alongside the PR #4 sandbox fix - there was never real concurrent access here, just an overly strict same-thread assertion tripping on a thread-identity change with only one logical caller. Verified: 244 backend tests pass (18 new for the parser + endpoint wiring + curry tool registration, 4 new for the TUI passthrough); the 12 pre-existing C/C++/Rust toolchain failures are unrelated and unchanged. Confirmed by hand over the real HTTP endpoint: successful dispatch, zero tool_request events (approval bypass working as designed), a format()-dunder exploit attempt still rejected by the vendored sandbox fix even through the new tool registration, malformed arguments rejected before ever reaching dispatch, and an unknown tool name rejected cleanly. Wheel rebuilt and content-checked (bin/check.sh's gate now also asserts slash_commands.py ships). Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
1789 lines
75 KiB
Python
1789 lines
75 KiB
Python
"""Tools a playbook can call during chat.
|
||
|
||
Ollama drives the calling: `/api/chat` with a `tools` param returns
|
||
`message.tool_calls`, and this module is just the registry + dispatch.
|
||
|
||
Most tools READ local state (memory, history, documents, models). Some act:
|
||
`web_search`/`fetch_url` make outbound HTTP requests, `remember` WRITES a
|
||
memory fact, `edit_playbook`/`edit_settings`/`edit_source` change the
|
||
assistant's own playbooks, settings, and (source checkout only) source code
|
||
(see synapse/self_edit.py), and `curry_*` reads and writes NexusOS's vendored
|
||
Curry ledger (see synapse/curry_store.py) — immutable versioned constants and
|
||
functions, with `curry_call_function` executing a previously declared one.
|
||
The per-playbook allowlist (`PlaybookItem.tools`) is the first gate — an
|
||
action tool only fires when a playbook explicitly lists it — and the
|
||
highest-risk tools in both families additionally always pause for per-call
|
||
approval regardless of the global action_tool_policy
|
||
(ALWAYS_ASK_ACTION_TOOLS, below). A message that's nothing but
|
||
`/tool_name(arg=val, ...)` skips that approval round-trip entirely and
|
||
dispatches directly — see synapse/slash_commands.py for why that's safe.
|
||
"""
|
||
from __future__ import annotations
|
||
|
||
import json
|
||
from typing import Awaitable, Callable
|
||
|
||
from . import code_run
|
||
from . import playbook_manager
|
||
from .curry_store import curry_db
|
||
from . import self_edit
|
||
from .memory.store import store, MemoryItem
|
||
from .ollama_manager import get_ollama_manager
|
||
|
||
|
||
async def _search_memory(query: str = "", **_) -> str:
|
||
q = (query or "").strip().lower()
|
||
hits = [
|
||
{"section": it.section, "text": it.text}
|
||
for it in store.all()
|
||
if not q
|
||
or q in it.text.lower()
|
||
or q in (it.section or "").lower()
|
||
or any(q in t.lower() for t in it.tags)
|
||
]
|
||
return json.dumps(hits[:20])
|
||
|
||
|
||
async def _search_history(query: str = "", **_) -> str:
|
||
# Hybrid recall: semantic (embeddings) unioned with lexical, falls back to
|
||
# lexical if embeddings are down. Same retrieval the chat endpoint uses.
|
||
convs = await store.semantic_search_conversations(
|
||
query or "", get_ollama_manager().embed, limit=3
|
||
)
|
||
return json.dumps([{"matches": c.get("matches", [])} for c in convs])
|
||
|
||
|
||
async def _list_models(**_) -> str:
|
||
return json.dumps(await get_ollama_manager().list_models())
|
||
|
||
|
||
async def _search_documents(query: str = "", **_) -> str:
|
||
hits = await store.search_documents(query or "", get_ollama_manager().embed, limit=3)
|
||
return json.dumps([{"title": h["title"], "text": h["text"]} for h in hits])
|
||
|
||
|
||
async def _get_time(**_) -> str:
|
||
from datetime import datetime
|
||
return json.dumps({"now": datetime.now().isoformat(timespec="seconds")})
|
||
|
||
|
||
async def _web_search(query: str = "", **_) -> str:
|
||
import asyncio as _a
|
||
from .search import web_search
|
||
res = await _a.to_thread(web_search, query or "", 4)
|
||
return res or "(no results)"
|
||
|
||
|
||
_FETCH_MAX_REDIRECTS = 5
|
||
|
||
|
||
def _ip_is_blocked(ip: str) -> bool:
|
||
"""True if an address is one an outbound fetch has no business reaching:
|
||
loopback, RFC1918/ULA private, link-local (incl. 169.254.169.254 cloud
|
||
metadata), multicast, reserved, or unspecified. IPv4-mapped IPv6 is unwrapped
|
||
first so ::ffff:127.0.0.1 can't sneak a loopback past the check."""
|
||
import ipaddress
|
||
try:
|
||
addr = ipaddress.ip_address(ip.split("%")[0]) # drop any IPv6 zone id
|
||
except ValueError:
|
||
return True # unparseable -> refuse rather than guess
|
||
mapped = getattr(addr, "ipv4_mapped", None)
|
||
if mapped is not None:
|
||
addr = mapped
|
||
return (
|
||
addr.is_loopback or addr.is_private or addr.is_link_local
|
||
or addr.is_multicast or addr.is_reserved or addr.is_unspecified
|
||
)
|
||
|
||
|
||
def _ssrf_guard(host: str) -> str | None:
|
||
"""Resolve a hostname and return an error string if ANY of its A/AAAA
|
||
records is a blocked address, else None. Checking every answer stops a name
|
||
from smuggling one private record alongside a public one.
|
||
|
||
ponytail: this validates then httpx re-resolves on connect, so a sub-second
|
||
DNS-rebind could still slip a private address through the TOCTOU gap. That's
|
||
an advanced attack against a playbook-gated, single-user tool; pin the
|
||
connection to the resolved IP if this ever faces untrusted callers."""
|
||
import socket
|
||
if not host:
|
||
return "missing host"
|
||
try:
|
||
infos = socket.getaddrinfo(host, None)
|
||
except socket.gaierror as e:
|
||
return f"cannot resolve host: {e}"
|
||
ips = {info[4][0] for info in infos}
|
||
if not ips:
|
||
return "host did not resolve"
|
||
blocked = [ip for ip in ips if _ip_is_blocked(ip)]
|
||
if blocked:
|
||
return f"refusing to fetch a private/loopback/link-local address ({', '.join(sorted(blocked))})"
|
||
return None
|
||
|
||
|
||
async def _fetch_url(url: str = "", **_) -> str:
|
||
import re
|
||
import httpx
|
||
from urllib.parse import urlparse, urljoin
|
||
url = (url or "").strip()
|
||
if not url.startswith(("http://", "https://")):
|
||
return json.dumps({"error": "url must start with http:// or https://"})
|
||
# SSRF guard: validate the host of the initial URL AND every redirect hop
|
||
# against the private/loopback/link-local block-list before connecting, so a
|
||
# granted fetch_url can't be steered at 127.0.0.1:11434, cloud metadata, or
|
||
# LAN hosts — and a public URL can't 302 its way there either.
|
||
try:
|
||
async with httpx.AsyncClient(timeout=15.0, follow_redirects=False) as c:
|
||
for _ in range(_FETCH_MAX_REDIRECTS + 1):
|
||
parsed = urlparse(url)
|
||
if parsed.scheme not in ("http", "https"):
|
||
return json.dumps({"error": "only http(s) URLs are allowed"})
|
||
err = _ssrf_guard(parsed.hostname or "")
|
||
if err:
|
||
return json.dumps({"error": f"blocked: {err}"})
|
||
r = await c.get(url, headers={"User-Agent": "NexusOS/1.0"})
|
||
location = r.headers.get("location")
|
||
if r.is_redirect and location:
|
||
url = urljoin(url, location)
|
||
continue
|
||
r.raise_for_status()
|
||
html = r.text
|
||
break
|
||
else:
|
||
return json.dumps({"error": "too many redirects"})
|
||
except Exception as e:
|
||
return json.dumps({"error": f"fetch failed: {e}"})
|
||
text = re.sub(r"(?is)<(script|style).*?</\1>", " ", html)
|
||
text = re.sub(r"(?s)<[^>]+>", " ", text)
|
||
text = re.sub(r"\s+", " ", text).strip()
|
||
return text[:4000]
|
||
|
||
|
||
async def _remember(text: str = "", section: str = "General", **_) -> str:
|
||
"""WRITE tool: persist a memory fact. First action tool — allowlist-gated."""
|
||
import uuid as _uuid
|
||
text = (text or "").strip()
|
||
if not text:
|
||
return json.dumps({"error": "text is required"})
|
||
store.add(MemoryItem(id=str(_uuid.uuid4()), section=(section or "General"), text=text))
|
||
return json.dumps({"saved": text, "section": section or "General"})
|
||
|
||
|
||
# Canvas drawing APIs a real visualization must use — resizing width/height alone
|
||
# clears the buffer and draws nothing (a failure mode small models hit often).
|
||
_CANVAS_DRAW_APIS = (
|
||
"fillrect", "strokerect", "filltext", "stroketext", "lineto", "arc(",
|
||
"beziercurveto", "quadraticcurveto", "fill(", "stroke(", "putimagedata",
|
||
"drawimage", "ellips(",
|
||
)
|
||
|
||
# Rejection threshold for a preview stage. Tiny 40×40 tiles (a recurring
|
||
# small-model collapse, often copied from earlier demos) can't show a sequence.
|
||
#
|
||
# These numbers are a threshold, never advice: every message that mentions a
|
||
# size quotes _STAGE_W/_STAGE_H instead. Weak models copy the first dimensions
|
||
# they read, so a message saying "at least 320x200, prefer 480x280" reliably
|
||
# produces 320x200 — three separate transcripts landed on exactly the minimum,
|
||
# including one that had a 480x280 example in front of it. Name one size.
|
||
_MIN_CANVAS_W = 320
|
||
_MIN_CANVAS_H = 200
|
||
|
||
# The size to ask for, and the only one any message should mention.
|
||
_STAGE_W = 480
|
||
_STAGE_H = 280
|
||
|
||
# Domain-agnostic interactive shell returned on reject as a *pattern* to adapt —
|
||
# not a finished demo for any particular algorithm. The model must implement
|
||
# generate() for the user's request (or ask them to clarify first).
|
||
_INTERACTIVE_SCAFFOLD_HTML = """<!DOCTYPE html>
|
||
<html><head><meta charset="utf-8"><style>
|
||
body{margin:0;font:14px/1.4 system-ui,sans-serif;background:#111;color:#eee;padding:12px}
|
||
.row{display:flex;gap:8px;align-items:center;margin-bottom:8px;flex-wrap:wrap}
|
||
input,button{font:inherit;padding:6px 10px}
|
||
canvas{display:block;width:480px;max-width:100%;height:auto;background:#1a1a1a;border:1px solid #333}
|
||
</style></head><body>
|
||
<div class="row">
|
||
<label>n <input id="n" type="number" min="1" value="20"></label>
|
||
<button id="go">Plot</button>
|
||
<span id="meta"></span>
|
||
</div>
|
||
<canvas id="c" width="480" height="280"></canvas>
|
||
<script>
|
||
const canvas = document.getElementById('c');
|
||
const ctx = canvas.getContext('2d');
|
||
|
||
/** Return an array of numbers (or {x,y} points) for THIS demo. */
|
||
function generate(n) {
|
||
// TODO: implement the user's algorithm / data here. Do not leave empty.
|
||
const seq = [];
|
||
for (let i = 0; i < n; i++) seq.push(i); // placeholder — replace
|
||
return seq;
|
||
}
|
||
|
||
function plot(seq) {
|
||
if (!seq || seq.length < 2) return;
|
||
const vals = seq.map(v => (typeof v === 'number' ? v : v.y));
|
||
const max = Math.max(1, ...vals);
|
||
const w = canvas.width, h = canvas.height, pad = 16;
|
||
ctx.clearRect(0, 0, w, h);
|
||
ctx.strokeStyle = '#3b82f6';
|
||
ctx.lineWidth = 2;
|
||
ctx.beginPath();
|
||
seq.forEach((v, i) => {
|
||
const x = pad + i * ((w - 2 * pad) / Math.max(1, seq.length - 1));
|
||
const y = h - pad - ((typeof v === 'number' ? v : v.y) / max) * (h - 2 * pad);
|
||
if (i === 0) ctx.moveTo(x, y); else ctx.lineTo(x, y);
|
||
});
|
||
ctx.stroke();
|
||
document.getElementById('meta').textContent = seq.length + ' points · max ' + max;
|
||
}
|
||
|
||
function go() {
|
||
plot(generate(+document.getElementById('n').value || 20));
|
||
}
|
||
document.getElementById('go').onclick = go;
|
||
go();
|
||
</script></body></html>"""
|
||
|
||
|
||
def _wants_data_visual(purpose: str = "", title: str = "", markup: str = "") -> bool:
|
||
"""True when the submission claims to be a chart/plot/interactive visual."""
|
||
blob = f"{purpose} {title} {markup}".lower()
|
||
return any(k in blob for k in (
|
||
"plot", "chart", "graph", "visual", "sequence", "orbit", "interactive",
|
||
"demo", "canvas", "diagram", "animation", "simulate", "conjecture",
|
||
))
|
||
|
||
|
||
# Component counterpart to _INTERACTIVE_SCAFFOLD_HTML, handed back when a
|
||
# jsx/tsx submission is rejected. Same contract: a pattern to adapt, not a demo
|
||
# to paste. No imports — the preview puts hooks and h/render in scope already,
|
||
# and there is no module loader in the sandbox to satisfy an import anyway.
|
||
_INTERACTIVE_SCAFFOLD_JSX = """export default function App() {
|
||
const canvasRef = useRef(null);
|
||
const [n, setN] = useState(20);
|
||
|
||
/** Return an array of numbers for THIS demo. */
|
||
function generate(count) {
|
||
// TODO: implement the user's algorithm / data here. Do not leave empty.
|
||
const seq = [];
|
||
for (let i = 0; i < count; i++) seq.push(i); // placeholder — replace
|
||
return seq;
|
||
}
|
||
|
||
useEffect(() => {
|
||
const canvas = canvasRef.current;
|
||
const ctx = canvas.getContext('2d');
|
||
const seq = generate(n);
|
||
const max = Math.max(1, ...seq);
|
||
const w = canvas.width, h = canvas.height, pad = 16;
|
||
ctx.clearRect(0, 0, w, h);
|
||
ctx.strokeStyle = '#3b82f6';
|
||
ctx.lineWidth = 2;
|
||
ctx.beginPath();
|
||
seq.forEach((v, i) => {
|
||
const x = pad + i * ((w - 2 * pad) / Math.max(1, seq.length - 1));
|
||
const y = h - pad - (v / max) * (h - 2 * pad);
|
||
if (i === 0) ctx.moveTo(x, y); else ctx.lineTo(x, y);
|
||
});
|
||
ctx.stroke();
|
||
}, [n]);
|
||
|
||
return (
|
||
<div style={{ font: '14px system-ui', background: '#111', color: '#eee', padding: 12 }}>
|
||
<label>n <input type="number" value={n} onInput={(e) => setN(+e.target.value || 1)} /></label>
|
||
<canvas ref={canvasRef} width="480" height="280" style={{ display: 'block', background: '#1a1a1a' }} />
|
||
</div>
|
||
);
|
||
}"""
|
||
|
||
|
||
def _wants_chart(purpose: str = "", title: str = "", markup: str = "") -> bool:
|
||
"""True only when the submission claims to draw *data* — a narrower test
|
||
than _wants_data_visual, which also counts "interactive" and "demo".
|
||
|
||
That wider net is right for an HTML fence, where an interactive demo with no
|
||
canvas is usually a model writing prose and calling it a visualization. It
|
||
is wrong for a component fence: a JSX counter or form is interactive through
|
||
its own elements and state, and demanding a <canvas> of it would reject the
|
||
most ordinary thing JSX is for."""
|
||
blob = f"{purpose} {title} {markup}".lower()
|
||
return any(k in blob for k in (
|
||
"plot", "chart", "graph", "visualiz", "diagram", "sequence", "orbit",
|
||
"histogram", "scatter",
|
||
))
|
||
|
||
|
||
def _decorative_svg_not_plot(markup: str) -> bool:
|
||
"""True when SVG is present but doesn't encode a multi-point data chart."""
|
||
import re
|
||
lower = markup.lower()
|
||
if "<svg" not in lower:
|
||
return False
|
||
# <defs> holds definitions, not output — nothing in it is drawn unless a
|
||
# <use>/fill references it. A long path parked in there was passing as proof
|
||
# of a real chart while the preview rendered an empty box.
|
||
if "<defs" in lower and not re.search(r"<use\b|url\(#", lower):
|
||
lower = re.sub(r"<defs\b.*?</defs\s*>", " ", lower, flags=re.S)
|
||
rich_poly = bool(re.search(
|
||
r"<polyline\b[^>]*\bpoints\s*=\s*[\"'][^\"']{40,}", lower,
|
||
))
|
||
rich_path = bool(re.search(
|
||
r"<path\b[^>]*\bd\s*=\s*[\"'][^\"']{40,}", lower,
|
||
))
|
||
builds_plot = bool(
|
||
re.search(r"createelementns\s*\(", lower)
|
||
and ("polyline" in lower or "path" in lower or "line" in lower)
|
||
and any(k in lower for k in ("foreach", "for (", "for(", "while(", "while ("))
|
||
and any(k in lower for k in ("seq", "points", "push(", "data"))
|
||
)
|
||
canvas_plot = "getcontext" in lower and any(a in lower for a in _CANVAS_DRAW_APIS)
|
||
return not (rich_poly or rich_path or builds_plot or canvas_plot)
|
||
|
||
|
||
def _critique_shared(markup: str) -> list[str]:
|
||
"""Checks that hold for every preview language."""
|
||
import re
|
||
issues: list[str] = []
|
||
lower = markup.lower()
|
||
|
||
external_attr = re.search(
|
||
r"""(?i)\b(?:src|srcset|href|xlink:href|poster|action|formaction|data)\b\s*=\s*(?:['"]\s*)?https?://""",
|
||
markup,
|
||
)
|
||
external_css = re.search(
|
||
r"""(?i)(?:url\s*\(\s*['"]?\s*https?://|@import\s+(?:url\s*\(\s*)?['"]?\s*https?://)""",
|
||
markup,
|
||
)
|
||
if external_attr or external_css:
|
||
issues.append(
|
||
"Remove external http(s) URLs — the sandboxed preview blocks them. "
|
||
"Inline CSS/JS; use data: URIs for images/fonts."
|
||
)
|
||
|
||
# Model talking about the chat UI instead of building the visual.
|
||
if any(p in lower for p in (
|
||
"preview/code", "render_preview", "live preview/code",
|
||
"paste the returned", "fenced block",
|
||
)):
|
||
issues.append(
|
||
"Do not describe the chat Preview UI — submit only the visualization "
|
||
"markup (canvas/SVG that plots data)."
|
||
)
|
||
|
||
return issues
|
||
|
||
|
||
def _critique_svg(markup: str, wants_plot: bool) -> list[str]:
|
||
import re
|
||
issues: list[str] = []
|
||
lower = markup.lower()
|
||
|
||
if "<svg" not in lower:
|
||
issues.append("SVG markup must include an <svg> root element.")
|
||
root = re.search(r"<svg\b[^>]*>", markup, re.I)
|
||
if root:
|
||
tag = root.group(0)
|
||
wm = re.search(r'\bwidth\s*=\s*["\']?(\d+)', tag, re.I)
|
||
hm = re.search(r'\bheight\s*=\s*["\']?(\d+)', tag, re.I)
|
||
if wm and int(wm.group(1)) < _MIN_CANVAS_W:
|
||
issues.append(
|
||
f'SVG width is {wm.group(1)}px — too small to read. Use '
|
||
f'width="{_STAGE_W}" height="{_STAGE_H}".'
|
||
)
|
||
if hm and int(hm.group(1)) < _MIN_CANVAS_H:
|
||
issues.append(
|
||
f'SVG height is {hm.group(1)}px — too small. Use height="{_STAGE_H}".'
|
||
)
|
||
if len(re.sub(r"\s+", "", markup)) < 60:
|
||
issues.append(
|
||
"SVG is too empty — add shapes (path/rect/circle/line/text) that "
|
||
"actually illustrate the idea."
|
||
)
|
||
if wants_plot and _decorative_svg_not_plot(markup):
|
||
issues.append(
|
||
"This SVG is decorative (gradient/rect/single line), not a data "
|
||
"chart. Build a <polyline>/<path> from many computed points, or "
|
||
"prefer a <canvas> with getContext + lineTo over an array."
|
||
)
|
||
return issues
|
||
|
||
|
||
def _critique_html(markup: str, wants_plot: bool) -> list[str]:
|
||
import re
|
||
issues: list[str] = []
|
||
lower = markup.lower()
|
||
|
||
if re.search(r"(?:width|height)\s*:\s*40px", markup, re.I):
|
||
issues.append(
|
||
"Do not use 40x40 CSS tiles — that is not a visualization. Size the "
|
||
f"stage {_STAGE_W}x{_STAGE_H}px."
|
||
)
|
||
|
||
if "<pre" in lower and ("<html" in lower or "<!doctype" in lower):
|
||
issues.append(
|
||
"Do not nest another HTML document inside <pre>. Put one interactive "
|
||
"<canvas> (or <svg>) in the body and draw there."
|
||
)
|
||
|
||
has_canvas = "<canvas" in lower
|
||
has_svg = "<svg" in lower
|
||
issues += _critique_prose(markup)
|
||
|
||
if wants_plot and not has_canvas and not has_svg:
|
||
issues.append(
|
||
"For chart/plot/interactive demos include a <canvas> or <svg> that "
|
||
"draws the data — prose alone is rejected."
|
||
)
|
||
if wants_plot and has_svg and not has_canvas and _decorative_svg_not_plot(markup):
|
||
issues.append(
|
||
"SVG stage present but it does not plot data (no multi-point "
|
||
"polyline/path, no JS that builds one from an array). Prefer "
|
||
f'<canvas width="{_STAGE_W}" height="{_STAGE_H}"> + getContext + lineTo.'
|
||
)
|
||
issues += _critique_canvas(markup)
|
||
|
||
if len(re.sub(r"\s+", "", markup)) < 40:
|
||
issues.append("markup is too short to be a useful preview.")
|
||
|
||
return issues
|
||
|
||
|
||
def _critique_prose(markup: str) -> list[str]:
|
||
"""Reject an explanation dressed up as a visualization — paragraphs, a list,
|
||
maybe a button that reveals more text, and nothing that draws.
|
||
|
||
Shared by html and jsx: a component returning four <p> elements is the same
|
||
non-answer as a page of them, and for a while jsx was accepted precisely
|
||
because this check lived only on the html side."""
|
||
import re
|
||
lower = markup.lower()
|
||
if "<canvas" in lower or "<svg" in lower:
|
||
return []
|
||
prose_tags = len(re.findall(r"<(?:p|li|h[1-6]|ul|ol)\b", lower))
|
||
toggle_only = prose_tags >= 3 and (
|
||
"display" in lower or "toggle" in lower or "<button" in lower
|
||
)
|
||
if prose_tags >= 4 or toggle_only:
|
||
return [
|
||
"This is an explanation, not a visualization. Draw the data on a "
|
||
f'<canvas width="{_STAGE_W}" height="{_STAGE_H}"> (or an <svg> chart) '
|
||
"— not paragraphs, lists, or a button that only reveals more text."
|
||
]
|
||
return []
|
||
|
||
|
||
def _critique_canvas(markup: str) -> list[str]:
|
||
"""Checks for markup that has a <canvas> in it, wherever that markup came
|
||
from — a plain HTML body or the JSX that renders one."""
|
||
import re
|
||
issues: list[str] = []
|
||
lower = markup.lower()
|
||
|
||
if "<canvas" in lower:
|
||
if "getcontext" not in lower:
|
||
issues.append(
|
||
"Canvas is present but never gets a 2D context — call "
|
||
"canvas.getContext('2d') and draw with it."
|
||
)
|
||
if not any(api in lower for api in _CANVAS_DRAW_APIS):
|
||
issues.append(
|
||
"Canvas never draws anything — plot each step with "
|
||
"fillRect/stroke/lineTo/arc/fillText (etc.). Do not only assign "
|
||
"canvas.width/height inside a loop; that clears the canvas."
|
||
)
|
||
|
||
for tag in re.findall(r"<canvas\b[^>]*>", markup, re.I):
|
||
wm = re.search(r'\bwidth\s*=\s*["\']?(\d+)', tag, re.I)
|
||
hm = re.search(r'\bheight\s*=\s*["\']?(\d+)', tag, re.I)
|
||
if wm and int(wm.group(1)) < _MIN_CANVAS_W:
|
||
issues.append(
|
||
f'Canvas width="{wm.group(1)}" is too small — use width="{_STAGE_W}" '
|
||
f'height="{_STAGE_H}", then map each data value to (x, y) pixels.'
|
||
)
|
||
if hm and int(hm.group(1)) < _MIN_CANVAS_H:
|
||
issues.append(
|
||
f'Canvas height="{hm.group(1)}" is too small — use height="{_STAGE_H}".'
|
||
)
|
||
if re.search(r'\bwidth\s*=\s*["\']?100%', tag, re.I):
|
||
issues.append(
|
||
"Use numeric canvas width/height attributes (e.g. width=\"480\"), "
|
||
"not percentages — the bitmap size must be explicit."
|
||
)
|
||
|
||
full_blit = bool(re.search(
|
||
r"(?:fillrect|clearrect)\s*\(\s*0\s*,\s*0\s*,\s*\d+\s*,\s*\d+\s*\)",
|
||
lower,
|
||
))
|
||
plots_points = bool(
|
||
re.search(r"lineto\s*\(", lower)
|
||
or re.search(r"fillrect\s*\(\s*(?!0\s*,\s*0)", lower)
|
||
or re.search(r"filltext\s*\(", lower)
|
||
or re.search(r"arc\s*\(", lower)
|
||
or re.search(r"strokerect\s*\(\s*(?!0\s*,\s*0)", lower)
|
||
)
|
||
if full_blit and not plots_points:
|
||
issues.append(
|
||
"You are only fillRect/clearRect(0,0,W,H) — that paints the whole "
|
||
"canvas, not the data. Collect values into an array, then for each "
|
||
"index i draw at x=i*step, y=height - value*scale (lineTo or "
|
||
"fillRect(x, y, barW, barH))."
|
||
)
|
||
|
||
return issues
|
||
|
||
|
||
def _critique_jsx(markup: str, wants_plot: bool) -> list[str]:
|
||
"""A JSX/TSX fence is one self-contained component. It is transformed and
|
||
mounted in the browser (interface/web/src/preview/), so the checks here are
|
||
the things that transform cannot recover from or would mount into nothing."""
|
||
import re
|
||
issues: list[str] = []
|
||
lower = markup.lower()
|
||
|
||
# Something has to be mounted: an explicit default export, a component named
|
||
# App, or some capitalized declaration to fall back to.
|
||
has_component = bool(
|
||
re.search(r"\bexport\s+default\b", markup)
|
||
or re.search(r"\bfunction\s+[A-Z]\w*", markup)
|
||
or re.search(r"\b(?:const|let|var)\s+[A-Z]\w*\s*=", markup)
|
||
or re.search(r"\bclass\s+[A-Z]\w*", markup)
|
||
)
|
||
if not has_component:
|
||
issues.append(
|
||
"No component to mount — define one with a capitalized name "
|
||
"(e.g. `function App() { ... }`) or `export default` it."
|
||
)
|
||
|
||
if "<" not in markup or not re.search(r"<[A-Za-z>]", markup):
|
||
issues.append(
|
||
"No JSX found — the component must return elements "
|
||
"(e.g. `return <div>…</div>;`)."
|
||
)
|
||
|
||
# Imports are stripped before the code runs: there is no module loader and
|
||
# no network in the sandbox. React/Preact itself is already in scope.
|
||
for module in re.findall(r"""\bfrom\s+['"]([^'"]+)['"]""", markup):
|
||
if module.split("/")[0] not in ("react", "react-dom", "preact"):
|
||
issues.append(
|
||
f"Cannot import '{module}' — the preview has no module loader and "
|
||
"no network. Inline what you need; React/Preact hooks are already "
|
||
"in scope without importing."
|
||
)
|
||
|
||
if wants_plot and "<canvas" not in lower and "<svg" not in lower:
|
||
issues.append(
|
||
"For chart/plot/interactive demos render a <canvas> or <svg> that "
|
||
"draws the data — prose alone is rejected."
|
||
)
|
||
|
||
issues += _critique_prose(markup)
|
||
issues += _critique_canvas(markup)
|
||
|
||
if len(re.sub(r"\s+", "", markup)) < 40:
|
||
issues.append("markup is too short to be a useful preview.")
|
||
|
||
return issues
|
||
|
||
|
||
# The one place that says which languages the render window supports. Each entry
|
||
# owns that language's validation; the tool schema's `lang` enum, the dispatch in
|
||
# _critique_render, and the capability line in the system prompt are all derived
|
||
# from these keys rather than repeating them.
|
||
#
|
||
# The frontend keeps its own matching registry (PREVIEW_LANGS in
|
||
# interface/web/src/Markdown.jsx) because the two sides need different things per
|
||
# language - this side validates, that side renders - and neither should depend
|
||
# on the other at runtime. tests/test_tools.py asserts the key sets stay equal,
|
||
# so drift fails the check gate instead of silently degrading to a plain code
|
||
# block in the chat.
|
||
PREVIEW_LANGS: dict[str, dict] = {
|
||
"html": {
|
||
"summary": "self-contained HTML document",
|
||
"critique": _critique_html,
|
||
# HTML also covers ordinary interactive UIs (forms, calculators, DOM
|
||
# demos). Only require a drawing surface when the request specifically
|
||
# claims to be a chart/plot/data visualization.
|
||
"wants_visual": _wants_chart,
|
||
"scaffold": _INTERACTIVE_SCAFFOLD_HTML,
|
||
},
|
||
"svg": {
|
||
"summary": "standalone SVG image",
|
||
"critique": _critique_svg,
|
||
"wants_visual": _wants_data_visual,
|
||
"scaffold": _INTERACTIVE_SCAFFOLD_HTML,
|
||
},
|
||
"jsx": {
|
||
"summary": "single Preact/React component (JSX)",
|
||
"critique": _critique_jsx,
|
||
"wants_visual": _wants_chart,
|
||
"scaffold": _INTERACTIVE_SCAFFOLD_JSX,
|
||
},
|
||
"tsx": {
|
||
"summary": "single Preact/React component (TypeScript JSX)",
|
||
"critique": _critique_jsx,
|
||
"wants_visual": _wants_chart,
|
||
"scaffold": _INTERACTIVE_SCAFFOLD_JSX,
|
||
},
|
||
}
|
||
|
||
|
||
def _scaffold_for(lang: str) -> str:
|
||
"""The starting pattern handed back on reject. Per-language: answering a
|
||
rejected component with a full HTML document tells the model to write the
|
||
wrong thing entirely."""
|
||
entry = PREVIEW_LANGS.get(lang)
|
||
return (entry["scaffold"] if entry else _INTERACTIVE_SCAFFOLD_HTML).strip()
|
||
|
||
|
||
def _lang_prose() -> str:
|
||
"""'html or svg' — the supported languages as a phrase for prompts/errors."""
|
||
names = list(PREVIEW_LANGS)
|
||
if len(names) < 2:
|
||
return names[0] if names else ""
|
||
return f"{', '.join(names[:-1])} or {names[-1]}"
|
||
|
||
|
||
def _critique_render(lang: str, markup: str, purpose: str = "", title: str = "") -> list[str]:
|
||
"""Cheap static checks so render_preview rejects empty/fake visuals before
|
||
the model pastes them into the chat as a 'working' demo.
|
||
|
||
Scope: things that RUN but are not what was asked for — a 40x40 stage, a
|
||
decorative gradient standing in for a chart, prose with no drawing in it, a
|
||
canvas that only paints itself one colour. These fail silently no matter
|
||
what, so static checks are the only thing that can catch them.
|
||
|
||
Not in scope: code that throws. The preview reports its own runtime errors
|
||
now (the bootstrap in interface/web/src/Markdown.jsx), so guessing at them
|
||
here bought nothing and cost accuracy. Three checks were removed once that
|
||
landed, each verified against the real error channel first:
|
||
|
||
const canvas = el.getContext('2d') ... ctx.lineTo()
|
||
-> "ReferenceError: ctx is not defined (line 4)"
|
||
function collatz() ... coll(27)
|
||
-> "ReferenceError: coll is not defined (line 5)"
|
||
document.createElementNS('line')
|
||
-> "TypeError: ... 2 arguments required, but only 1 present. (line 3)"
|
||
|
||
The real messages are better than the regexes were: they carry a line
|
||
number, and they catch *any* undefined name rather than the two spellings
|
||
someone thought to anticipate. Resist re-adding a static check for anything
|
||
that already throws."""
|
||
entry = PREVIEW_LANGS.get(lang)
|
||
if entry is None:
|
||
return [f"unsupported preview language {lang!r} — use {_lang_prose()}."]
|
||
# What counts as "claimed a visual" differs by language: see _wants_chart.
|
||
wants_plot = entry["wants_visual"](purpose, title, markup)
|
||
return _critique_shared(markup) + entry["critique"](markup, wants_plot)
|
||
|
||
|
||
def _with_scaffold(payload: dict, lang: str, attempt: int) -> dict:
|
||
"""Attach the starting pattern, but only from the second rejection on.
|
||
|
||
A complete, styled, runnable document handed to a struggling model does not
|
||
get adapted — it gets pasted, and then it persists in the conversation and
|
||
comes back as retrieved context for the next request, carrying its example
|
||
domain with it. Transcripts show exactly that: a scaffold's CSS reappearing
|
||
verbatim in an answer to an unrelated prompt, in a conversation where this
|
||
tool was never even called. So the first rejection says only what is wrong;
|
||
the pattern appears once that has not been enough."""
|
||
if attempt < 1:
|
||
return payload
|
||
return {
|
||
**payload,
|
||
"scaffold": _scaffold_for(lang),
|
||
"scaffold_note": (
|
||
"A pattern to adapt, not an answer to paste. Replace generate() with "
|
||
"the logic this request actually needs; keep nothing you do not use."
|
||
),
|
||
}
|
||
|
||
|
||
async def _render_preview(
|
||
lang: str = "html",
|
||
title: str = "",
|
||
markup: str = "",
|
||
purpose: str = "",
|
||
_attempt: int = 0,
|
||
**_,
|
||
) -> str:
|
||
"""Validate + package a live-preview fence. Read-only: nothing is executed
|
||
server-side; the chat UI renders the returned fence in a sandboxed iframe.
|
||
|
||
`_attempt` is supplied by the tool loop, not by the model — it is how many
|
||
times this call has already been rejected in the current turn."""
|
||
lang = (lang or "html").strip().lower()
|
||
markup = (markup or "").strip()
|
||
title = (title or "").strip()
|
||
purpose = (purpose or "").strip()
|
||
|
||
if lang not in PREVIEW_LANGS:
|
||
return json.dumps({"ok": False, "error": f"lang must be {_lang_prose()}"})
|
||
if not markup:
|
||
return json.dumps(_with_scaffold({
|
||
"ok": False,
|
||
"error": (
|
||
f"markup is required — send the complete {lang} for the visual you "
|
||
"want, with all CSS and JS inline and no external URLs."
|
||
),
|
||
}, lang, _attempt))
|
||
|
||
issues = _critique_render(lang, markup, purpose=purpose, title=title)
|
||
if issues:
|
||
return json.dumps(_with_scaffold({
|
||
"ok": False,
|
||
"issues": issues,
|
||
"hint": (
|
||
"Fix these and call render_preview again, building the thing that "
|
||
f"was actually asked for. Draw on a {_STAGE_W}x{_STAGE_H} stage; "
|
||
"compute the values into an array first, then plot them point by "
|
||
"point with lineTo/fillRect(x,y,w,h); add <input>/<button> controls "
|
||
"when it should be interactive."
|
||
),
|
||
"purpose": purpose or None,
|
||
}, lang, _attempt))
|
||
|
||
fence = f"```{lang}\n{markup}\n```"
|
||
return json.dumps({
|
||
"ok": True,
|
||
"title": title or None,
|
||
"purpose": purpose or None,
|
||
"instruction": (
|
||
"Write a short intro, then paste this fenced block exactly as it is. "
|
||
"Do not wrap it in a second fence, resize it, or rewrite the code."
|
||
),
|
||
"fence": fence,
|
||
})
|
||
|
||
|
||
# ---------------------------------------------------------------------------
|
||
# run_snippet — the execution track
|
||
# ---------------------------------------------------------------------------
|
||
#
|
||
# render_preview and run_snippet are deliberately separate tools over separate
|
||
# registries. A preview is validated here and *rendered* by the browser inside a
|
||
# sandboxed frame; a snippet is *executed* on the host by synapse/code_run.py and
|
||
# comes back as terminal output. Stretching one tool across both would have meant
|
||
# a `lang` enum where half the values run server-side and half don't, and a
|
||
# single description that could only be vague about which. The split is the
|
||
# feature: the model picks a track by picking a tool.
|
||
#
|
||
# Result fence: one ```nexus-run block whose body is JSON (source + streams +
|
||
# exit code), so the code and the output it produced cannot be separated by a
|
||
# model pasting only half of it. Backticks in the source are re-encoded as \u0060
|
||
# — still valid JSON, and it cannot terminate the fence early.
|
||
_RUN_FENCE_LANG = "nexus-run"
|
||
|
||
|
||
def _run_fence(payload: dict) -> str:
|
||
body = json.dumps(payload, ensure_ascii=False).replace("`", "\\u0060")
|
||
return f"```{_RUN_FENCE_LANG}\n{body}\n```"
|
||
|
||
|
||
def _run_lang_prose() -> str:
|
||
return code_run.lang_prose()
|
||
|
||
|
||
def _run_scaffold_for(lang: str) -> str:
|
||
"""Minimal entry-point patterns for a struggling model. Shown from the
|
||
second rejection on — same discipline as render_preview's scaffold."""
|
||
return {
|
||
"python": "print(sum(range(10)))\n",
|
||
"c": (
|
||
"#include <stdio.h>\n"
|
||
"int main(void) {\n"
|
||
' printf("%d\\n", 42);\n'
|
||
" return 0;\n"
|
||
"}\n"
|
||
),
|
||
"cpp": (
|
||
"#include <iostream>\n"
|
||
"int main() {\n"
|
||
' std::cout << 42 << "\\n";\n'
|
||
" return 0;\n"
|
||
"}\n"
|
||
),
|
||
"rust": 'fn main() { println!("{}", 42); }\n',
|
||
"erlang": 'main(_) -> io:format("~p~n", [42]).\n',
|
||
}.get(lang, f"(a short self-contained {lang} program that prints to stdout)\n")
|
||
|
||
|
||
def _with_run_scaffold(payload: dict, lang: str, attempt: int) -> dict:
|
||
"""Attach a starting pattern from the second rejection on. First rejection
|
||
stays issues-only so a pasteable scaffold does not become the answer."""
|
||
if attempt < 1:
|
||
return payload
|
||
return {
|
||
**payload,
|
||
"scaffold": _run_scaffold_for(lang),
|
||
"scaffold_note": (
|
||
"A pattern to adapt, not an answer to paste. Keep the entry point, "
|
||
"print results to stdout, and drop anything you do not use."
|
||
),
|
||
}
|
||
|
||
|
||
async def _run_snippet(
|
||
lang: str = "",
|
||
source: str = "",
|
||
stdin: str = "",
|
||
title: str = "",
|
||
_attempt: int = 0,
|
||
**_,
|
||
) -> str:
|
||
"""Compile and run a snippet, returning a fence that carries both the source
|
||
and what it printed.
|
||
|
||
ACTION tool: it executes code on this machine. `action_tool_policy` gates it
|
||
(withheld on "off", per-call Approve/Deny on "ask"), and synapse/code_run.py
|
||
documents exactly how much containment the run itself gets — which is less
|
||
than the word "sandbox" would imply.
|
||
|
||
`_attempt` is supplied by the tool loop, not by the model — it is how many
|
||
times this call has already been rejected in the current turn."""
|
||
import asyncio as _a
|
||
|
||
key = code_run.resolve_lang(lang)
|
||
source = (source or "").strip()
|
||
title = (title or "").strip()
|
||
|
||
if key not in code_run.RUN_LANGS:
|
||
return json.dumps({
|
||
"ok": False,
|
||
"error": f"lang must be {_run_lang_prose()} (got {lang!r}). "
|
||
"For HTML, SVG or JSX use render_preview instead — those are "
|
||
"rendered in the browser, not executed here.",
|
||
})
|
||
if not source:
|
||
return json.dumps(_with_run_scaffold({
|
||
"ok": False,
|
||
"error": f"source is required — send the complete {key} program, "
|
||
"entry point included.",
|
||
}, key, _attempt))
|
||
|
||
issues = code_run.critique(key, source)
|
||
if issues:
|
||
return json.dumps(_with_run_scaffold({
|
||
"ok": False,
|
||
"issues": issues,
|
||
"hint": (
|
||
"Fix these and call run_snippet again. The program runs in a throwaway "
|
||
"directory with no network and a few seconds of CPU: no downloads, no "
|
||
"absolute paths, no unbounded loops. Print your results to stdout."
|
||
),
|
||
}, key, _attempt))
|
||
|
||
result = await _a.to_thread(code_run.run, key, source, stdin or "")
|
||
if not result.get("ok"):
|
||
# A failed compile hands back the compiler's own diagnostics: they name
|
||
# the line and the fix, and paraphrasing them here would only lose that.
|
||
return json.dumps({k: v for k, v in result.items() if v not in ("", None)})
|
||
|
||
return json.dumps({
|
||
"ok": True,
|
||
"lang": key,
|
||
"exit_code": result["exit_code"],
|
||
"stdout": result["stdout"],
|
||
"stderr": result["stderr"],
|
||
"title": title or f"{key} output",
|
||
"instruction": (
|
||
"This ran for real — the output below is what it printed. Write a short "
|
||
"intro, then paste the fenced block exactly as it is. Do not wrap it in a "
|
||
"second fence, retype the output, or claim results it does not show."
|
||
),
|
||
"fence": _run_fence({
|
||
"lang": key,
|
||
"source": source,
|
||
"stdout": result["stdout"],
|
||
"stderr": result["stderr"],
|
||
"exit_code": result["exit_code"],
|
||
}),
|
||
})
|
||
|
||
|
||
# Result fence for the three self-edit tools: same "carry the change and what
|
||
# happened to it in one block" idea as _run_fence, so the model can't paste a
|
||
# claimed diff that doesn't match what was actually applied.
|
||
_EDIT_FENCE_LANG = "nexus-edit"
|
||
|
||
|
||
def _edit_fence(payload: dict) -> str:
|
||
body = json.dumps(payload, ensure_ascii=False).replace("`", "\\u0060")
|
||
return f"```{_EDIT_FENCE_LANG}\n{body}\n```"
|
||
|
||
|
||
async def _edit_source(path: str = "", new_content: str = "", summary: str = "", **_) -> str:
|
||
"""ACTION tool: writes a file in the live project tree and commits it. See
|
||
synapse/self_edit.py for the boundary check, the size cap, and exactly what
|
||
the git commit does and doesn't guarantee."""
|
||
import asyncio as _a
|
||
|
||
result = await _a.to_thread(self_edit.apply_source_edit, path, new_content or "", summary or "")
|
||
if result.get("ok"):
|
||
result["instruction"] = (
|
||
"This was written and committed for real. Paste the fence unchanged, then "
|
||
"tell the user plainly that a restart is needed for it to take effect — "
|
||
"this file is not reloaded into the running process."
|
||
)
|
||
result["fence"] = _edit_fence({"kind": "source", **result})
|
||
return json.dumps(result)
|
||
|
||
|
||
async def _edit_playbook(
|
||
id: str = "", title: str = "", goal: str = "", instructions: str = "",
|
||
tags: list | None = None, tools: list | None = None, model: str = "",
|
||
make_active: bool = False, **_,
|
||
) -> str:
|
||
"""ACTION tool: create or update a playbook. Fields left unset keep their
|
||
current value — this merges, it does not replace. make_active is a
|
||
separate, explicit flag: without it, an edit can never accidentally become
|
||
the active system prompt."""
|
||
try:
|
||
result = playbook_manager.persist_playbook(
|
||
{
|
||
"id": id, "title": title, "goal": goal, "instructions": instructions,
|
||
"tags": tags, "tools": tools, "model": model,
|
||
},
|
||
merge=True,
|
||
)
|
||
except ValueError as e:
|
||
return json.dumps({"ok": False, "error": str(e)})
|
||
if make_active:
|
||
playbook_manager.make_main(result["id"])
|
||
result["is_main_playbook"] = True
|
||
result["ok"] = True
|
||
result["fence"] = _edit_fence({"kind": "playbook", **result})
|
||
return json.dumps(result)
|
||
|
||
|
||
async def _edit_settings(changes: dict | None = None, **_) -> str:
|
||
"""ACTION tool: change one or more runtime settings. Unknown keys are
|
||
silently ignored, exactly like PUT /settings already does."""
|
||
changes = changes if isinstance(changes, dict) else {}
|
||
if not changes:
|
||
return json.dumps({"ok": False, "error": "changes must be a non-empty object"})
|
||
preview = self_edit.preview_settings_edit({"changes": changes})
|
||
if not preview.get("ok"):
|
||
return json.dumps(preview)
|
||
if not preview.get("applied"):
|
||
return json.dumps({
|
||
"ok": False,
|
||
"error": "no recognized settings keys in changes",
|
||
"ignored_unknown": preview.get("ignored_unknown", []),
|
||
})
|
||
store.update_settings({k: v["after"] for k, v in preview["applied"].items()})
|
||
preview["ok"] = True
|
||
preview["fence"] = _edit_fence({"kind": "settings", **preview})
|
||
return json.dumps(preview)
|
||
|
||
|
||
# Curry (synapse/curry_core.py, vendored) — NexusOS's immutable, versioned
|
||
# fact store. Its own methods raise (KeyError/ValueError/TypeError/RuntimeError)
|
||
# on the failures a caller should see as a normal, expected result rather than
|
||
# a crash (unknown id, version conflict, retired reference, etc.) — dispatch()
|
||
# would already catch anything unhandled, but that produces a generic
|
||
# "toolname failed: ..." string instead of the {"ok": False, "error": ...}
|
||
# shape every other tool in this file returns, so it's caught locally here too.
|
||
_CURRY_FENCE_LANG = "nexus-curry"
|
||
|
||
|
||
def _curry_fence(payload: dict) -> str:
|
||
body = json.dumps(payload, ensure_ascii=False, default=str).replace("`", "\\u0060")
|
||
return f"```{_CURRY_FENCE_LANG}\n{body}\n```"
|
||
|
||
|
||
async def _curry_call(fn, *args, **kwargs) -> dict:
|
||
# Not run.to_thread()'d like run_snippet/self_edit's blocking work: curry_db
|
||
# holds one sqlite3 connection for its whole lifetime (unlike
|
||
# PersistentMemoryStore, which opens/closes a fresh one per call), and
|
||
# sqlite3 forbids using a connection from any thread but the one that
|
||
# created it. curry_db is created at import time on the same thread this
|
||
# runs on (the asyncio event loop thread), so calling it directly here is
|
||
# both correct and, for local-file SQLite, fast enough not to need
|
||
# offloading anyway.
|
||
try:
|
||
result = fn(*args, **kwargs)
|
||
return {"ok": True, "result": result}
|
||
except (KeyError, ValueError, TypeError, RuntimeError) as e:
|
||
return {"ok": False, "error": str(e)}
|
||
|
||
|
||
async def _curry_declare_constant(
|
||
id: str = "", version: int = 0, value=None, type_signature: str = "",
|
||
description: str = "", **_,
|
||
) -> str:
|
||
"""ACTION tool: declare a new, immutable version of a named constant."""
|
||
out = await _curry_call(
|
||
curry_db.declare_constant, id, version, value, type_signature, description or None
|
||
)
|
||
if out["ok"]:
|
||
out = {"ok": True, "id": id, "version": version}
|
||
out["fence"] = _curry_fence({"kind": "declare_constant", **out})
|
||
return json.dumps(out)
|
||
|
||
|
||
async def _curry_get_constant(id: str = "", version: int = 0, **_) -> str:
|
||
"""Retrieve a constant by exact id and version."""
|
||
return json.dumps(await _curry_call(curry_db.get_constant, id, version))
|
||
|
||
|
||
async def _curry_get_constant_latest(id: str = "", **_) -> str:
|
||
"""Retrieve the most recent active (non-retired) version of a constant."""
|
||
return json.dumps(await _curry_call(curry_db.get_constant_latest, id))
|
||
|
||
|
||
async def _curry_list_constants(active_only: bool = True, **_) -> str:
|
||
"""List all declared constants and their latest versions."""
|
||
return json.dumps(await _curry_call(curry_db.list_constants, active_only))
|
||
|
||
|
||
async def _curry_retire_constant(id: str = "", version: int = 0, reason: str = "", **_) -> str:
|
||
"""ACTION tool: retire (tombstone) a constant version. Does not delete it —
|
||
the version stays readable by exact id+version, just excluded from
|
||
"latest" lookups and blocked from new declarations that depend on it."""
|
||
out = await _curry_call(
|
||
curry_db.retire_constant_with_reason, id, version, reason or "retired via tool call"
|
||
)
|
||
return json.dumps(out)
|
||
|
||
|
||
async def _curry_declare_function(
|
||
name: str = "", version: int = 0, body: str = "",
|
||
constant_bindings: dict | None = None, function_bindings: dict | None = None,
|
||
is_pure: bool = False, expected_args: list | None = None,
|
||
description: str = "", arg_descriptions: dict | None = None, **_,
|
||
) -> str:
|
||
"""ACTION tool: declare a new, immutable version of a named function. Body
|
||
is a single Python expression (no statements) over stdlib-only builtins,
|
||
checked by curry_core.py's own static validator before this ever runs —
|
||
but that validator is a tripwire against habitual mistakes, not a
|
||
security boundary; treat it the same as run_snippet's containment."""
|
||
out = await _curry_call(
|
||
curry_db.declare_function, name, version, body,
|
||
constant_bindings or {}, function_bindings or {}, is_pure,
|
||
expected_args, description or None, arg_descriptions,
|
||
)
|
||
if out["ok"]:
|
||
out = {"ok": True, "name": name, "version": version}
|
||
out["fence"] = _curry_fence({"kind": "declare_function", **out})
|
||
return json.dumps(out)
|
||
|
||
|
||
async def _curry_get_function(name: str = "", version: int = 0, **_) -> str:
|
||
"""Retrieve a function definition by exact name and version."""
|
||
return json.dumps(await _curry_call(curry_db.get_function, name, version))
|
||
|
||
|
||
async def _curry_list_functions(active_only: bool = True, **_) -> str:
|
||
"""List all declared functions and their latest versions."""
|
||
return json.dumps(await _curry_call(curry_db.list_functions, active_only))
|
||
|
||
|
||
async def _curry_call_function(name: str = "", version: int = 0, args: dict | None = None, **_) -> str:
|
||
"""ACTION tool: execute a previously declared function version with the
|
||
given runtime arguments. Locked constant/function dependencies resolve
|
||
automatically; pure functions are memoized."""
|
||
out = await _curry_call(curry_db.call_function, name, version, args or {})
|
||
if out["ok"]:
|
||
out["fence"] = _curry_fence({"kind": "call_function", "name": name, "version": version, **out})
|
||
return json.dumps(out)
|
||
|
||
|
||
async def _curry_retire_function(name: str = "", version: int = 0, reason: str = "", **_) -> str:
|
||
"""ACTION tool: retire (tombstone) a function version. Does not delete it."""
|
||
out = await _curry_call(
|
||
curry_db.retire_function_with_reason, name, version, reason or "retired via tool call"
|
||
)
|
||
return json.dumps(out)
|
||
|
||
|
||
# name -> (schema, callable). Schema is the OpenAI/Ollama function-tool format.
|
||
REGISTRY: dict[str, tuple[dict, Callable[..., Awaitable[str]]]] = {
|
||
"search_memory": (
|
||
{
|
||
"type": "function",
|
||
"function": {
|
||
"name": "search_memory",
|
||
"description": "Search the user's persistent memory facts. Empty query returns all facts.",
|
||
"parameters": {
|
||
"type": "object",
|
||
"properties": {"query": {"type": "string", "description": "text to match"}},
|
||
},
|
||
},
|
||
},
|
||
_search_memory,
|
||
),
|
||
"search_history": (
|
||
{
|
||
"type": "function",
|
||
"function": {
|
||
"name": "search_history",
|
||
"description": "Search past conversations for exchanges containing the query text.",
|
||
"parameters": {
|
||
"type": "object",
|
||
"properties": {"query": {"type": "string"}},
|
||
"required": ["query"],
|
||
},
|
||
},
|
||
},
|
||
_search_history,
|
||
),
|
||
"list_models": (
|
||
{
|
||
"type": "function",
|
||
"function": {
|
||
"name": "list_models",
|
||
"description": "List the locally installed Ollama models.",
|
||
"parameters": {"type": "object", "properties": {}},
|
||
},
|
||
},
|
||
_list_models,
|
||
),
|
||
"search_documents": (
|
||
{
|
||
"type": "function",
|
||
"function": {
|
||
"name": "search_documents",
|
||
"description": "Search the user's uploaded documents for relevant passages.",
|
||
"parameters": {
|
||
"type": "object",
|
||
"properties": {"query": {"type": "string"}},
|
||
"required": ["query"],
|
||
},
|
||
},
|
||
},
|
||
_search_documents,
|
||
),
|
||
"get_time": (
|
||
{
|
||
"type": "function",
|
||
"function": {
|
||
"name": "get_time",
|
||
"description": "Get the current local date and time.",
|
||
"parameters": {"type": "object", "properties": {}},
|
||
},
|
||
},
|
||
_get_time,
|
||
),
|
||
"render_preview": (
|
||
{
|
||
"type": "function",
|
||
"function": {
|
||
"name": "render_preview",
|
||
# Written as instructions TO you, imperative and short. Earlier
|
||
# versions narrated what "the user" wants and listed numbered
|
||
# requirements; weak models echoed that narration back as their
|
||
# reply — asking the user to clarify an already-clear request,
|
||
# in the third person, instead of building anything. Keep this
|
||
# terse, keep it second-person, and add nothing the model can
|
||
# recite in place of acting.
|
||
"description": (
|
||
f"Build a working visual — chart, plot, diagram, interactive demo — "
|
||
f"as self-contained {_lang_prose()} and send it here to check. "
|
||
f"Draw on a {_STAGE_W}x{_STAGE_H} stage. Compute your values into an "
|
||
"array, then plot them point by point (canvas: getContext, then "
|
||
"lineTo/fillRect(x,y,w,h)/arc per point). Add <input>/<button> "
|
||
"controls if it should be interactive. Inline all CSS and JS; the "
|
||
"preview is sandboxed with no network, so external URLs will not "
|
||
"load. Build what was asked for, not a similar demo you know better. "
|
||
"Rejected: fix what `issues` lists and send it again. "
|
||
"Accepted: paste the returned `fence` into your reply unchanged."
|
||
),
|
||
"parameters": {
|
||
"type": "object",
|
||
"properties": {
|
||
"lang": {
|
||
"type": "string",
|
||
"enum": list(PREVIEW_LANGS),
|
||
"description": (
|
||
"Preview language tag for the fenced block: "
|
||
+ "; ".join(
|
||
f"{name} ({spec['summary']})"
|
||
for name, spec in PREVIEW_LANGS.items()
|
||
)
|
||
),
|
||
},
|
||
"title": {
|
||
"type": "string",
|
||
"description": "Short label for the visual.",
|
||
},
|
||
"purpose": {
|
||
"type": "string",
|
||
"description": "One sentence: what this visual shows.",
|
||
},
|
||
"markup": {
|
||
"type": "string",
|
||
"description": (
|
||
"Full self-contained HTML document or SVG. Inline all "
|
||
"CSS/JS. No external script/style/img URLs."
|
||
),
|
||
},
|
||
},
|
||
"required": ["lang", "markup"],
|
||
},
|
||
},
|
||
},
|
||
_render_preview,
|
||
),
|
||
"run_snippet": (
|
||
{
|
||
"type": "function",
|
||
"function": {
|
||
"name": "run_snippet",
|
||
# Same house style as render_preview: imperative, second person,
|
||
# nothing the model can recite back in place of acting.
|
||
"description": (
|
||
f"Compile and run a short {code_run.lang_prose()} program on this "
|
||
"machine and get its real output back. Use this when the answer "
|
||
"depends on what the code actually does — output, a computed "
|
||
"result, whether it compiles. Write one self-contained file with "
|
||
"its entry point; there is no package manager, no network, a "
|
||
f"throwaway working directory and {code_run.RUN_TIMEOUT:g}s of "
|
||
"runtime, so bound your loops and print results to stdout. "
|
||
"For HTML, SVG or JSX use render_preview instead. "
|
||
"Rejected: fix what `issues` or `stderr` says and send it again. "
|
||
"Accepted: paste the returned `fence` into your reply unchanged, "
|
||
"and describe only the output it actually contains."
|
||
),
|
||
"parameters": {
|
||
"type": "object",
|
||
"properties": {
|
||
"lang": {
|
||
"type": "string",
|
||
"enum": list(code_run.RUN_LANGS),
|
||
"description": (
|
||
"Language to run: "
|
||
+ "; ".join(
|
||
f"{name} ({spec['summary']})"
|
||
for name, spec in code_run.RUN_LANGS.items()
|
||
)
|
||
),
|
||
},
|
||
"title": {
|
||
"type": "string",
|
||
"description": "Short label for what this program does.",
|
||
},
|
||
"source": {
|
||
"type": "string",
|
||
"description": (
|
||
"The complete single-file program, entry point included. "
|
||
"Standard library only."
|
||
),
|
||
},
|
||
"stdin": {
|
||
"type": "string",
|
||
"description": "Optional text piped to the program's stdin.",
|
||
},
|
||
},
|
||
"required": ["lang", "source"],
|
||
},
|
||
},
|
||
},
|
||
_run_snippet,
|
||
),
|
||
"web_search": (
|
||
{
|
||
"type": "function",
|
||
"function": {
|
||
"name": "web_search",
|
||
"description": "Search the web (DuckDuckGo) and return the top result snippets.",
|
||
"parameters": {
|
||
"type": "object",
|
||
"properties": {"query": {"type": "string"}},
|
||
"required": ["query"],
|
||
},
|
||
},
|
||
},
|
||
_web_search,
|
||
),
|
||
"fetch_url": (
|
||
{
|
||
"type": "function",
|
||
"function": {
|
||
"name": "fetch_url",
|
||
"description": "Fetch a web page and return its visible text (truncated).",
|
||
"parameters": {
|
||
"type": "object",
|
||
"properties": {"url": {"type": "string", "description": "http(s) URL"}},
|
||
"required": ["url"],
|
||
},
|
||
},
|
||
},
|
||
_fetch_url,
|
||
),
|
||
"remember": (
|
||
{
|
||
"type": "function",
|
||
"function": {
|
||
"name": "remember",
|
||
"description": "Save a durable fact to the user's persistent memory.",
|
||
"parameters": {
|
||
"type": "object",
|
||
"properties": {
|
||
"text": {"type": "string", "description": "the fact to remember"},
|
||
"section": {"type": "string", "description": "optional category, e.g. Health"},
|
||
},
|
||
"required": ["text"],
|
||
},
|
||
},
|
||
},
|
||
_remember,
|
||
),
|
||
"edit_source": (
|
||
{
|
||
"type": "function",
|
||
"function": {
|
||
"name": "edit_source",
|
||
"description": (
|
||
"Rewrite a file in this project's own source tree and commit the "
|
||
"change. Requires human approval every time — the person reviews a "
|
||
"real diff before anything is written. Send the COMPLETE new file "
|
||
"content, not a patch; the server computes the diff itself. `path` "
|
||
"is relative to the project root (e.g. \"synapse/tools.py\"), never "
|
||
"absolute. Only available in a source checkout, not a packaged "
|
||
"install. Writing the file does not restart the running process — "
|
||
"say so plainly once it's applied."
|
||
),
|
||
"parameters": {
|
||
"type": "object",
|
||
"properties": {
|
||
"path": {
|
||
"type": "string",
|
||
"description": "Project-relative path to the file, e.g. synapse/tools.py",
|
||
},
|
||
"new_content": {
|
||
"type": "string",
|
||
"description": "The complete replacement content of the file.",
|
||
},
|
||
"summary": {
|
||
"type": "string",
|
||
"description": "One line describing the change, used as the commit message.",
|
||
},
|
||
},
|
||
"required": ["path", "new_content"],
|
||
},
|
||
},
|
||
},
|
||
_edit_source,
|
||
),
|
||
"edit_playbook": (
|
||
{
|
||
"type": "function",
|
||
"function": {
|
||
"name": "edit_playbook",
|
||
"description": (
|
||
"Create or update a playbook — the instructions that shape how the "
|
||
"assistant behaves. Requires human approval every time. Fields you "
|
||
"omit keep their current value; this merges into the existing "
|
||
"playbook, it does not replace it. Set make_active=true only when "
|
||
"this playbook should become the active system prompt — never as a "
|
||
"side effect of an ordinary edit. Omit `id` to create a new playbook."
|
||
),
|
||
"parameters": {
|
||
"type": "object",
|
||
"properties": {
|
||
"id": {"type": "string", "description": "Existing playbook id to update; omit to create new."},
|
||
"title": {"type": "string", "description": "Short name for the playbook."},
|
||
"goal": {"type": "string", "description": "One-line statement of what this playbook is for."},
|
||
"instructions": {"type": "string", "description": "The actual instructions/system prompt text."},
|
||
"tags": {"type": "array", "items": {"type": "string"}, "description": "Routing tags."},
|
||
"tools": {"type": "array", "items": {"type": "string"}, "description": "Tool names this playbook grants."},
|
||
"model": {"type": "string", "description": "Preferred Ollama model for this playbook, or blank for auto."},
|
||
"make_active": {
|
||
"type": "boolean",
|
||
"description": "Set true to make this the active system prompt. Default false.",
|
||
},
|
||
},
|
||
"required": [],
|
||
},
|
||
},
|
||
},
|
||
_edit_playbook,
|
||
),
|
||
"edit_settings": (
|
||
{
|
||
"type": "function",
|
||
"function": {
|
||
"name": "edit_settings",
|
||
"description": (
|
||
"Change one or more runtime settings (e.g. model, temperature, "
|
||
"action_tool_policy, memory_model). Requires human approval every "
|
||
"time. Send only the keys you actually want to change — unrecognized "
|
||
"keys are silently ignored, and existing values are left untouched."
|
||
),
|
||
"parameters": {
|
||
"type": "object",
|
||
"properties": {
|
||
"changes": {
|
||
"type": "object",
|
||
"description": "Partial map of setting name to new value.",
|
||
},
|
||
},
|
||
"required": ["changes"],
|
||
},
|
||
},
|
||
},
|
||
_edit_settings,
|
||
),
|
||
"curry_declare_constant": (
|
||
{
|
||
"type": "function",
|
||
"function": {
|
||
"name": "curry_declare_constant",
|
||
"description": (
|
||
"Declare a new, immutable version of a named constant in the Curry "
|
||
"ledger. Requires human approval every time. `version` must exceed "
|
||
"the constant's current max version — versions are append-only, "
|
||
"never overwritten. type_signature is one of Float64, Int32, "
|
||
"String, Blob, Json, Tokens, Currency, Bool."
|
||
),
|
||
"parameters": {
|
||
"type": "object",
|
||
"properties": {
|
||
"id": {"type": "string", "description": "Constant identifier."},
|
||
"version": {"type": "integer", "description": "Must exceed the current max version for this id."},
|
||
"value": {"description": "The value to store, matching type_signature."},
|
||
"type_signature": {"type": "string", "description": "Float64 | Int32 | String | Blob | Json | Tokens | Currency | Bool"},
|
||
"description": {"type": "string", "description": "What this constant means and why this value."},
|
||
},
|
||
"required": ["id", "version", "value", "type_signature"],
|
||
},
|
||
},
|
||
},
|
||
_curry_declare_constant,
|
||
),
|
||
"curry_get_constant": (
|
||
{
|
||
"type": "function",
|
||
"function": {
|
||
"name": "curry_get_constant",
|
||
"description": "Retrieve a Curry constant by its exact id and version.",
|
||
"parameters": {
|
||
"type": "object",
|
||
"properties": {
|
||
"id": {"type": "string", "description": "Constant identifier."},
|
||
"version": {"type": "integer", "description": "Exact version to retrieve."},
|
||
},
|
||
"required": ["id", "version"],
|
||
},
|
||
},
|
||
},
|
||
_curry_get_constant,
|
||
),
|
||
"curry_get_constant_latest": (
|
||
{
|
||
"type": "function",
|
||
"function": {
|
||
"name": "curry_get_constant_latest",
|
||
"description": "Retrieve the most recent active (non-retired) version of a Curry constant.",
|
||
"parameters": {
|
||
"type": "object",
|
||
"properties": {
|
||
"id": {"type": "string", "description": "Constant identifier."},
|
||
},
|
||
"required": ["id"],
|
||
},
|
||
},
|
||
},
|
||
_curry_get_constant_latest,
|
||
),
|
||
"curry_list_constants": (
|
||
{
|
||
"type": "function",
|
||
"function": {
|
||
"name": "curry_list_constants",
|
||
"description": "List every constant declared in the Curry ledger.",
|
||
"parameters": {
|
||
"type": "object",
|
||
"properties": {
|
||
"active_only": {"type": "boolean", "description": "If true (default), exclude retired constants."},
|
||
},
|
||
"required": [],
|
||
},
|
||
},
|
||
},
|
||
_curry_list_constants,
|
||
),
|
||
"curry_retire_constant": (
|
||
{
|
||
"type": "function",
|
||
"function": {
|
||
"name": "curry_retire_constant",
|
||
"description": (
|
||
"Retire (tombstone) a Curry constant version. Requires human approval "
|
||
"every time. This does not delete anything — the version stays "
|
||
"readable by exact id+version, it's just excluded from 'latest' "
|
||
"lookups going forward."
|
||
),
|
||
"parameters": {
|
||
"type": "object",
|
||
"properties": {
|
||
"id": {"type": "string", "description": "Constant identifier."},
|
||
"version": {"type": "integer", "description": "Version to retire."},
|
||
"reason": {"type": "string", "description": "Why this version is being retired."},
|
||
},
|
||
"required": ["id", "version"],
|
||
},
|
||
},
|
||
},
|
||
_curry_retire_constant,
|
||
),
|
||
"curry_declare_function": (
|
||
{
|
||
"type": "function",
|
||
"function": {
|
||
"name": "curry_declare_function",
|
||
"description": (
|
||
"Declare a new, immutable version of a named function in the Curry "
|
||
"ledger. Requires human approval every time. body is a SINGLE Python "
|
||
"expression (no statements, no imports) over stdlib-only builtins — "
|
||
"reference bound constants/functions by name via constant_bindings / "
|
||
"function_bindings, and any additional runtime arguments via "
|
||
"expected_args. `version` must exceed the function's current max "
|
||
"version."
|
||
),
|
||
"parameters": {
|
||
"type": "object",
|
||
"properties": {
|
||
"name": {"type": "string", "description": "Function name."},
|
||
"version": {"type": "integer", "description": "Must exceed the current max version for this name."},
|
||
"body": {"type": "string", "description": "Single Python expression, e.g. \"amount * (1 + rate)\"."},
|
||
"constant_bindings": {"type": "object", "description": "Dict mapping constant id to the exact version to bind, e.g. {\"rate\": 1}."},
|
||
"function_bindings": {"type": "object", "description": "Dict mapping nested function name to the exact version to bind."},
|
||
"is_pure": {"type": "boolean", "description": "If true, results are memoized in the execution cache."},
|
||
"expected_args": {"type": "array", "items": {"type": "string"}, "description": "Runtime argument names the caller must supply to curry_call_function."},
|
||
"description": {"type": "string", "description": "What this function computes and which constants it binds."},
|
||
"arg_descriptions": {"type": "object", "description": "Per-argument hint strings, e.g. {\"amount\": \"USD, e.g. 100.00\"}."},
|
||
},
|
||
"required": ["name", "version", "body"],
|
||
},
|
||
},
|
||
},
|
||
_curry_declare_function,
|
||
),
|
||
"curry_get_function": (
|
||
{
|
||
"type": "function",
|
||
"function": {
|
||
"name": "curry_get_function",
|
||
"description": "Retrieve a Curry function definition by its exact name and version.",
|
||
"parameters": {
|
||
"type": "object",
|
||
"properties": {
|
||
"name": {"type": "string", "description": "Function name."},
|
||
"version": {"type": "integer", "description": "Exact version to retrieve."},
|
||
},
|
||
"required": ["name", "version"],
|
||
},
|
||
},
|
||
},
|
||
_curry_get_function,
|
||
),
|
||
"curry_list_functions": (
|
||
{
|
||
"type": "function",
|
||
"function": {
|
||
"name": "curry_list_functions",
|
||
"description": "List every function declared in the Curry ledger, including expected_args for building curry_call_function calls.",
|
||
"parameters": {
|
||
"type": "object",
|
||
"properties": {
|
||
"active_only": {"type": "boolean", "description": "If true (default), exclude retired functions."},
|
||
},
|
||
"required": [],
|
||
},
|
||
},
|
||
},
|
||
_curry_list_functions,
|
||
),
|
||
"curry_call_function": (
|
||
{
|
||
"type": "function",
|
||
"function": {
|
||
"name": "curry_call_function",
|
||
"description": (
|
||
"Execute a previously declared Curry function version with runtime "
|
||
"arguments. Requires human approval every time. Use "
|
||
"curry_list_functions or curry_get_function first to discover "
|
||
"expected_args."
|
||
),
|
||
"parameters": {
|
||
"type": "object",
|
||
"properties": {
|
||
"name": {"type": "string", "description": "Function name."},
|
||
"version": {"type": "integer", "description": "Exact version to execute."},
|
||
"args": {"type": "object", "description": "Runtime arguments as a flat dict, e.g. {\"amount\": 100}."},
|
||
},
|
||
"required": ["name", "version"],
|
||
},
|
||
},
|
||
},
|
||
_curry_call_function,
|
||
),
|
||
"curry_retire_function": (
|
||
{
|
||
"type": "function",
|
||
"function": {
|
||
"name": "curry_retire_function",
|
||
"description": (
|
||
"Retire (tombstone) a Curry function version. Requires human approval "
|
||
"every time. Does not delete anything."
|
||
),
|
||
"parameters": {
|
||
"type": "object",
|
||
"properties": {
|
||
"name": {"type": "string", "description": "Function name."},
|
||
"version": {"type": "integer", "description": "Version to retire."},
|
||
"reason": {"type": "string", "description": "Why this version is being retired."},
|
||
},
|
||
"required": ["name", "version"],
|
||
},
|
||
},
|
||
},
|
||
_curry_retire_function,
|
||
),
|
||
}
|
||
|
||
|
||
# Tools that act (write local state or reach the network). These require an
|
||
# explicit consent gate (settings.allow_action_tools) on top of the per-playbook
|
||
# allowlist — a playbook granting one isn't enough on its own. The highest-risk
|
||
# tools in the self-edit and curry families additionally always pause for
|
||
# per-call approval regardless of that global policy — see
|
||
# ALWAYS_ASK_ACTION_TOOLS and chat.py. curry_call_function is an action
|
||
# because it executes code (a declared function body), the same reasoning
|
||
# that makes run_snippet an action tool despite not writing to any ledger.
|
||
ACTION_TOOLS = frozenset({
|
||
"web_search", "fetch_url", "remember", "run_snippet",
|
||
"edit_playbook", "edit_settings", "edit_source",
|
||
"curry_declare_constant", "curry_retire_constant",
|
||
"curry_declare_function", "curry_retire_function", "curry_call_function",
|
||
})
|
||
|
||
# Curry write/execute tools that always pause for per-call approval regardless
|
||
# of the global action_tool_policy, on the same reasoning as
|
||
# self_edit.ALWAYS_ASK_TOOLS: a policy of "allow" set for convenience on an
|
||
# unrelated tool must never silently also unlock unattended ledger writes or
|
||
# code execution. Read-only curry_get_*/curry_list_* tools are not action
|
||
# tools at all and are unaffected.
|
||
CURRY_ALWAYS_ASK_TOOLS = frozenset({
|
||
"curry_declare_constant", "curry_retire_constant",
|
||
"curry_declare_function", "curry_retire_function", "curry_call_function",
|
||
})
|
||
|
||
# The union chat.py actually checks — one place, so a future tool family
|
||
# doesn't have to remember there are two sets to update.
|
||
ALWAYS_ASK_ACTION_TOOLS = self_edit.ALWAYS_ASK_TOOLS | CURRY_ALWAYS_ASK_TOOLS
|
||
|
||
# Action tools offered on a *cue* rather than only via a playbook allowlist —
|
||
# the run track is a standing UI capability like the render window, but unlike
|
||
# render_preview it executes code, so it stays behind the action gate. Listed
|
||
# here so main.py can advertise it on a "run this" without a playbook edit.
|
||
CUED_ACTION_TOOLS = frozenset({"run_snippet"})
|
||
|
||
# Always advertised when the user asks for a visual (see wants_render_preview).
|
||
# Not playbook-gated — the render window is a standing UI capability.
|
||
STANDING_TOOLS = frozenset({"render_preview"})
|
||
|
||
# User-message cues that justify running the (slow, non-stream) tool loop with
|
||
# render_preview. Kept narrow so ordinary chat isn't blocked behind a tool turn.
|
||
_RENDER_HINTS = (
|
||
"visual", "visuals", "visualize", "visualization", "chart", "charts",
|
||
"graph", "graphs", "diagram", "diagrams", "canvas", "plot", "plots",
|
||
"interactive", "animation", "animations", "render_preview",
|
||
"render preview", "svg", "draw me", "live preview",
|
||
"demonstrate", "demo", "html demo", "html snippet", "html file",
|
||
# Ways of asking for something that reacts to the pointer. "interactive"
|
||
# alone missed "mouse-over sensitive", and with it the whole feature.
|
||
"hover", "mouse", "drag", "click on", "real-time", "realtime",
|
||
"simulation", "simulations", "simulate", "particle", "particles", "animate",
|
||
# Every language the render window can display. Naming one is asking for a
|
||
# preview, and this way a language added to PREVIEW_LANGS starts hinting
|
||
# for itself instead of being unreachable until someone edits this tuple -
|
||
# which is exactly what happened to jsx/tsx.
|
||
) + tuple(PREVIEW_LANGS)
|
||
|
||
|
||
# Cues for run_snippet. Unlike _RENDER_HINTS these are verb phrases, not topic
|
||
# words, and deliberately do not include the language names: "write me a python
|
||
# function" is not a request to execute anything, and putting `python` in here
|
||
# would drag every mention of the language into a slow non-stream tool round.
|
||
# Asking for the *output* is the signal, so that is what these match.
|
||
_RUN_HINTS = (
|
||
"run this", "run it", "run that", "run the code", "run this code",
|
||
"run my code", "run the program", "run and show", "actually run",
|
||
"run snippet", "run_snippet", "execute this", "execute it", "execute the code",
|
||
"compile", "compiles", "does it compile", "does this compile",
|
||
"show the output", "show me the output", "what does it print",
|
||
"what does this print", "what's the output", "whats the output",
|
||
"what is the output", "actual output", "real output",
|
||
)
|
||
|
||
|
||
def _mentions(message: str, hints: tuple[str, ...]) -> bool:
|
||
"""True if any hint appears in `message` as a whole word/phrase.
|
||
|
||
The lookarounds rather than \\b: several hints end in a non-word character
|
||
("what's the output"), where \\b would anchor to the apostrophe instead of
|
||
the phrase and match inside longer words.
|
||
"""
|
||
import re
|
||
lower = (message or "").lower()
|
||
return any(
|
||
re.search(rf"(?<![A-Za-z0-9_]){re.escape(hint)}(?![A-Za-z0-9_])", lower)
|
||
for hint in hints
|
||
)
|
||
|
||
|
||
def wants_render_preview(message: str) -> bool:
|
||
"""True when this turn should advertise render_preview / enter the tool loop."""
|
||
return _mentions(message, _RENDER_HINTS)
|
||
|
||
|
||
def wants_code_run(message: str) -> bool:
|
||
"""True when this turn is asking for code to actually be executed.
|
||
|
||
Only a hint: run_snippet is an action tool, so this can never be what makes
|
||
it available — `action_tool_policy` still has to be off "off" first.
|
||
"""
|
||
return _mentions(message, _RUN_HINTS)
|
||
|
||
|
||
def is_action(name: str) -> bool:
|
||
return name in ACTION_TOOLS
|
||
|
||
|
||
def schemas_for(names: list[str], allow_actions: bool = True) -> list[dict]:
|
||
"""Tool schemas for a playbook's allowlist; unknown names are dropped.
|
||
When allow_actions is False, action tools are withheld so the model can't
|
||
even call them."""
|
||
return [
|
||
REGISTRY[n][0] for n in (names or [])
|
||
if n in REGISTRY and (allow_actions or not is_action(n))
|
||
]
|
||
|
||
|
||
def standing_schemas() -> list[dict]:
|
||
"""Schemas that ship with visual turns (currently just render_preview)."""
|
||
return schemas_for(sorted(STANDING_TOOLS), allow_actions=True)
|
||
|
||
|
||
def run_schemas() -> list[dict]:
|
||
"""Schemas that ship with "run this" turns (currently just run_snippet).
|
||
|
||
Callers must have already established that actions are permitted — these are
|
||
action tools and schemas_for would happily hand them over regardless.
|
||
"""
|
||
return schemas_for(sorted(CUED_ACTION_TOOLS), allow_actions=True)
|
||
|
||
|
||
async def dispatch(name: str, args: dict | None) -> str:
|
||
"""Run a tool by name. Never raises — returns an error string on failure."""
|
||
entry = REGISTRY.get(name)
|
||
if not entry:
|
||
return json.dumps({"error": f"unknown tool: {name}"})
|
||
try:
|
||
return await entry[1](**(args or {}))
|
||
except Exception as e: # a broken tool must not kill the chat loop
|
||
return json.dumps({"error": f"{name} failed: {e}"})
|
GitNexus · part of the Nexus stack