Raven-1/tests/test_plugin_runtime_binding.py

184 lines
6.6 KiB
Python
Raw Permalink Normal View History

fix(tools): an array sent as its own JSON is still an array ## Summary Reported as "deliver_files is broken". It is not: **the tool never ran.** Every `deliver_files` call on one install was refused before dispatch -- five in a row in a single session -- and nothing was ever delivered: ``` Error: Invalid parameters for tool 'deliver_files': files should be array ``` What the model sent was the array, spelled as a string: ```json {"files": "[{\"path\": \"...poster.png\", \"title\": \"...\"}]", "message": "..."} ``` `cast_params` already forgives that spelling for every scalar type -- `"3"` for an integer, `"true"` for a boolean -- and did not for arrays or objects, which are exactly the types a model is most likely to stringify. ## Type - [x] Fix - [ ] Feature ## The edge of the forgiving is the design When the value is a string, try `json.loads`, and take the result **only if it is the type the schema asked for**. A string that is not JSON, or is JSON of the wrong shape, is returned unchanged so the validator still names the real problem -- the point is to widen what is accepted, not to swallow a type error. **Nothing here can raise.** Turning "your argument has the wrong type" into a traceback would be worse than the refusal it replaces, so the parse is wrapped and failure falls through. **A `type: "string"` field is never parsed.** JSON is ordinary text in a message body or a file's content, and replacing it would hand the tool something it never asked for. Two independent things prevent it today (the string branch returns before the parse is reached, and the parse only wins on a type match), which is why the case pinning it is a tripwire rather than a load-bearing assertion -- said so in its own docstring rather than left to look like coverage. ## Not just deliver_files Six built-in tools declare an array or object parameter and were all exposed: | tool | parameter | |---|---| | `deliver_files` | `files` | | `ask_user` | `questions` | | `image_generate` | `images` | | `message` | `media` | | `create_playbook` | `skills` | | `video_generate` | `params` | The same failure was already seen from the other side: `acp_client/acp_agent.py` records `update_task_state` and `operations should be array` from a real dispatch. ## Verification ``` make lint-python ruff check + ruff format --check, 1850 files make lint-imports 9 contracts kept, 0 broken pytest tests/ 19,699 passed, 7 failed (all pre-existing, see below) ``` - [x] Relevant tests pass locally - [x] Relevant lint / type checks pass locally - [ ] User-facing docs or screenshots -- no user-facing string **End to end, through the real tool**, with the exact payload that failed five times in production: ``` files sent as: str -> [{"path": ".../poster.png", "title": "..."}] returned: Delivered 1 file(s): poster.png (72 B). still "should be array": False registered in the store: True ``` Twelve cases, five mutations, each killing at least one: | mutation | red | |---|---| | drop the whole branch (the old behaviour) | 3 | | raise instead of falling through on a parse failure | 5 | | accept the parse whatever type it yielded | 1 | | do not re-cast the parsed value, so items are not cast | 1 | | widen the tuple to include `string` | **survived -- unreachable** | The last is worth stating rather than hiding: that mutation cannot change behaviour, because the `type == "string"` branch returns earlier and the type check would refuse the result anyway. I also tried moving the new block above that branch, alone and combined with widening the tuple, and neither reddens anything either. The invariant holds by construction; the case that documents it says so. **Pre-existing failures, checked against the base.** 7 failed in the full run, all in `tests/test_security_network.py` (`test_a_redirect_to_a_dangerous_address_is_refused` and siblings). Running that file alone on this same tree gives 4 -- the count is order-dependent -- and the change touches only `raven/agent/tools/params.py` plus a new test file, neither of which `raven/security/network.py` imports. The failures are the box's resolver: it returns poisoned answers roughly half the time, which I measured separately (`en.wikipedia.org` came back as `2001::1` on 3 of 6 lookups). Not introduced here, and worth someone's attention on its own. ## Risk - [x] Security impact considered -- widens what reaches a tool's own validation, not what bypasses it. A parsed value is still validated against the same schema, and a parse that yields the wrong type is refused exactly as before - [x] Backward compatibility considered -- a call that worked before takes the identical path; the new branch is reached only by a `str` where an array or object is declared, which previously always failed - [x] Rollback path is clear -- single-commit revert ## Related Issues N/A ## Swept `mr-review-patterns` pre-submit sweep, over `git diff origin/main...HEAD`. - **1 (a write location moved)** -- nothing moved; a branch was added to an existing dispatch on `target_type`. Grepped every caller of `cast_params`: the registry is the only one, and it runs `validate_params` right after, which is the reader this affects. - **2 (a value with several entry points)** -- the cast is the single funnel every tool argument passes through, which is why the fix belongs here and not in `deliver_files`. - **3 (a guard)** -- checked both directions: what it wrongly lets through (JSON of the wrong type, non-JSON text -- both refused, with cases) and what it wrongly refuses (a real list, unaffected, with a case). - **4 (tests)** -- five mutations above, one honestly reported as unreachable. Failing-ID sets diffed against the base rather than counts. - **6 (stale text)** -- the module docstring enumerated what is coerced ("a string that should be an integer"); it now names arrays and objects too. Grepped the tracked tree for other statements of the rule: none. - **9 (error handling)** -- the new path adds one way to end and it is caught: `json.loads` raises `ValueError` on bad text and `TypeError` on a non-string, and both fall through to the untouched value. - **11** -- `git ls-files | grep -c node_modules` is 0; the commit-message gate was run, not eyeballed. Not applicable, with reasons: **3b, 3c, 3d, 3e** (no tool schema produced, no state moved, one lane, no assertion lost its symbol); **4b, 4c, 4d** (no platform branch, no shell, no rendered number); **5, 5b** (no dependency; `json` is stdlib and already imported across this package); **6b, 6c** (no backup, no generated artifact); **7, 7b** (nothing on disk); **8** (no asset, no secret); **10** (no prompt text). ## Deliberately not fixed The error text still ends with `[Analyze the error above and try a different approach.]`, and the model plainly did not -- it sent the same shape five times. A message naming the expected spelling (`files must be a JSON array, not a string containing one`) would give it something to act on. That is a change to the registry's error surface, affects every tool, and deserves its own argument. ## Moved from !594 Opened on a compliant source branch. AGENTS.md 2.1 asks for a 3-5 word snake_case description and the previous branch had seven; GitLab cannot change an existing merge request's source branch, so the change moves rather than the name. Same commits, rebased onto `c2a9768f`. Both code findings from the review on !594 are already in them -- `RecursionError` caught with its regression case, and the implementation comment reduced to the two non-obvious reasons. Nothing else changed, so that review carries.
2026-09-09 15:42:36 +00:00
"""Late binding: the contribution shape for tools that need the living loop.
A factory runs in the assembly root, before the loop exists. These pin the
whole grant path at the loop level: a declarer receives the frozen handles
exactly once with the loop's own objects in them, a non-declarer is left
alone, and a tool that raises while binding is unregistered loudly instead of
serving half-bound.
"""
from __future__ import annotations
import pytest
from raven.agent.loop import AgentLoop
from raven.agent.loop.bundles import EngineWiring, HostWiring, ToolWiring, TurnPolicy
from raven.contracts.llm_provider import LLMResponse
from raven.plugins.context import RuntimeHandles
from raven.providers.base import LLMProvider
class _Provider(LLMProvider):
def __init__(self) -> None:
super().__init__(api_key="test")
async def chat(
self,
messages,
tools=None,
model=None,
max_tokens=4096,
temperature=0.7,
reasoning_effort=None,
tool_choice=None,
):
return LLMResponse(content="ok", finish_reason="stop")
def get_default_model(self) -> str:
return "fake/default"
class _Bindable:
name = "bindable_probe"
description = "records the handles the loop grants"
parameters = {"type": "object", "properties": {}}
def __init__(self) -> None:
self.bound: list[RuntimeHandles] = []
def bind_runtime(self, handles: RuntimeHandles) -> None:
self.bound.append(handles)
async def execute(self, **kwargs) -> str:
return "ok"
class _Plain:
name = "plain_probe"
description = "declares nothing and must be left alone"
parameters = {"type": "object", "properties": {}}
async def execute(self, **kwargs) -> str:
return "ok"
class _Explosive:
name = "explosive_probe"
description = "raises while binding and must not serve half-bound"
parameters = {"type": "object", "properties": {}}
def bind_runtime(self, handles: RuntimeHandles) -> None:
raise RuntimeError("no thanks")
async def execute(self, **kwargs) -> str:
return "ok"
def _loop(tmp_path, plugin_tools, playbook_config=None) -> AgentLoop:
loop = AgentLoop(
provider=_Provider(),
workspace=tmp_path,
model="fake/model",
policy=TurnPolicy(max_iterations=2),
host=HostWiring(),
tools=ToolWiring(restrict_to_workspace=True, plugin_tools=plugin_tools),
engine=EngineWiring(playbook_config=playbook_config),
)
async def _noop(**_kw) -> None:
return None
loop._start_executor = _noop
loop._connect_mcp = _noop
return loop
def test_a_declarer_is_bound_once_with_the_loops_own_handles(tmp_path):
probe = _Bindable()
plain = _Plain()
loop = _loop(tmp_path, [probe, plain])
assert len(probe.bound) == 1
handles = probe.bound[0]
assert handles.subagent_registry is loop.subagents.registry
assert handles.session_dir == loop.sessions.session_dir
assert handles.subagents_paused() == loop.subagents.paused
assert loop.tools.get("plain_probe") is plain, "a non-declarer is registered untouched"
with pytest.raises(Exception):
handles.session_dir = None # type: ignore[misc]
def test_a_tool_that_raises_while_binding_is_unregistered(tmp_path):
loop = _loop(tmp_path, [_Explosive(), _Plain()])
assert loop.tools.get("explosive_probe") is None, "half-bound tools must not serve"
assert loop.tools.get("plain_probe") is not None, "one bad plugin does not take the others down"
def test_the_playbook_funnel_reaches_the_bundled_tools_by_binding(tmp_path):
"""The whole grant path for the shelf's first real resident: the loop
assembles the funnel, and the two bundled tools serve exactly that object
-- same runtime, same generator, same store, same adopt."""
from raven.agent.tools.create_playbook import CreatePlaybookTool
from raven.agent.tools.load_playbook import LoadPlaybookTool
from raven.config.schema import PlaybookConfig
load, create = LoadPlaybookTool(), CreatePlaybookTool()
loop = _loop(tmp_path, [load, create], playbook_config=PlaybookConfig(enabled=True))
assert loop._playbooks is not None, "the loop was configured to build the funnel"
assert loop.tools.get("load_playbook") is load
assert loop.tools.get("create_playbook") is create
assert load._runtime is loop._playbooks
assert create._generator is loop._playbooks.generator
assert create._store is loop._playbooks.store
assert create._adopt == loop._playbooks.adopt
def test_a_loop_without_playbooks_takes_the_bundled_tools_off_the_table(tmp_path):
"""No funnel, no service: the tools decline and are unregistered quietly,
so `playbooks.enabled: false` turns the feature exactly as far off as it
was when the loop registered the tools itself."""
from raven.agent.tools.create_playbook import CreatePlaybookTool
from raven.agent.tools.load_playbook import LoadPlaybookTool
loop = _loop(tmp_path, [LoadPlaybookTool(), CreatePlaybookTool(), _Plain()])
assert loop.tools.get("load_playbook") is None, "an unbound loader must not serve"
assert loop.tools.get("create_playbook") is None
assert loop.tools.get("plain_probe") is not None, "a decline does not take the others down"
def test_the_wake_grant_is_namespaced_by_the_build_stamp(tmp_path):
"""[seam-1] The loop mints the wake grant per tool from the identity the
plugin build stamped -- a namespace the plugin chose for itself would be a
namespace it could steal -- and a tool without a stamp, or a host without
a scheduler, gets None."""
import time
from raven.proactive_engine.schedulers.cron.service import CronService
stamped = _Bindable()
stamped.contributed_by = "plug-a"
bare = _Bindable()
bare.name = "bindable_bare"
cron = CronService(tmp_path / "jobs.json", allowed_channels=None)
loop = _loop(tmp_path, [stamped, bare])
assert stamped.bound[0].wake_scheduler is None, "no scheduler on this host, no grant"
loop2 = AgentLoop(
provider=_Provider(),
workspace=tmp_path,
model="fake/model",
policy=TurnPolicy(max_iterations=2),
host=HostWiring(cron_service=cron),
tools=ToolWiring(restrict_to_workspace=True, plugin_tools=[stamped, bare]),
)
granted = stamped.bound[-1].wake_scheduler
assert granted is not None
job = granted.schedule_wake("c1", int(time.time() * 1000) + 60_000, "look", channel="tui")
assert job.id == "wake:plug-a:c1"
assert bare.bound[-1].wake_scheduler is None, "no stamped identity, no namespace to grant into"
assert loop2 is not None