Skip to content

armgit / strix

Public Imported from GitHub
0

Open-source AI penetration testing tool to find and fix your app’s vulnerabilities.

fix: address audit findings — SDK plumbing, TUI bus, dead code

0xallam572ef2a2afd7
diff --git a/pyproject.toml b/pyproject.toml
index 73dfb8e7..f6a49c1b 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -35,12 +35,10 @@ classifiers = [
dependencies = [
"litellm[proxy]>=1.83.0",
"openai-agents[litellm]==0.14.6",
- "tenacity>=9.0.0",
"pydantic[email]>=2.11.3",
"rich",
"docker>=7.1.0",
"textual>=6.0.0",
- "xmltodict>=0.13.0",
"requests>=2.32.0",
"cvss>=3.2",
"traceloop-sdk>=0.53.0",
@@ -119,7 +117,6 @@ pretty = true
[[tool.mypy.overrides]]
module = [
"litellm.*",
- "tenacity.*",
"numpydoc.*",
"rich.*",
"IPython.*",
@@ -294,7 +291,6 @@ ignore = [
"strix/tools/web_search/tool.py" = ["TC002"]
"strix/tools/file_edit/tools.py" = ["TC002"]
"strix/tools/reporting/tool.py" = ["TC002"]
-"strix/tools/load_skill/tool.py" = ["TC002"]
"strix/tools/finish/tool.py" = ["TC002"]
"strix/tools/browser/tool.py" = ["TC002"]
"strix/tools/terminal/tool.py" = ["TC002"]
@@ -420,7 +416,7 @@ force_grid_wrap = 0
use_parentheses = true
ensure_newline_before_comments = true
known_first_party = ["strix"]
-known_third_party = ["fastapi", "pydantic", "litellm", "tenacity"]
+known_third_party = ["fastapi", "pydantic", "litellm"]
# ============================================================================
# Pytest Configuration
diff --git a/strix/agents/factory.py b/strix/agents/factory.py
index 0fc77d8a..8455e5c0 100644
--- a/strix/agents/factory.py
+++ b/strix/agents/factory.py
@@ -21,8 +21,9 @@ Two flavors:
Caido tools come from ``CaidoCapability.tools()`` automatically via
the SDK's capability merge — we don't include them here. Skills are
-injected via the prompt; the model can also load more at runtime via
-the ``load_skill`` tool.
+injected via the prompt at scan-bring-up time; runtime skill loading
+isn't exposed as a tool any more (the legacy implementation reached
+into a global agent registry that no longer exists).
References:
- PLAYBOOK.md §4.3 (graph tool wiring)
@@ -54,7 +55,6 @@ from strix.tools.file_edit.tools import (
str_replace_editor,
)
from strix.tools.finish.tool import finish_scan
-from strix.tools.load_skill.tool import load_skill
from strix.tools.notes.tools import (
create_note,
delete_note,
@@ -109,8 +109,6 @@ _BASE_TOOLS: tuple[Tool, ...] = (
search_files,
# Reporting
create_vulnerability_report,
- # Skill loading
- load_skill,
# Sandbox primitives
browser_action,
terminal_execute,
@@ -140,8 +138,7 @@ def build_strix_agent(
name: Agent name. Surfaces in traces and the bus's ``names`` map.
Defaults to ``"strix"`` for the root; create_agent passes
distinct names per child.
- skills: Skills to preload into the system prompt. The agent can
- also load more at runtime via the ``load_skill`` tool.
+ skills: Skills to preload into the system prompt.
is_root: Selects the tool list and ``tool_use_behavior``.
Root carries ``finish_scan`` and stops there; child carries
``agent_finish`` and stops there.
diff --git a/strix/agents/prompt.py b/strix/agents/prompt.py
index 795deaa3..9ace5550 100644
--- a/strix/agents/prompt.py
+++ b/strix/agents/prompt.py
@@ -67,9 +67,7 @@ def render_system_prompt(
"""Render the system prompt.
Args:
- skills: Skills the caller wants preloaded into the prompt
- context (the agent can also load more at runtime via the
- ``load_skill`` tool).
+ skills: Skills the caller wants preloaded into the prompt context.
scan_mode: ``"deep" | "fast" | ...``. Maps to ``scan_modes/<mode>``
skill.
is_whitebox: When True, the source-aware whitebox skill stack
diff --git a/strix/agents/prompts/system_prompt.jinja b/strix/agents/prompts/system_prompt.jinja
index 8c89ef24..a67a29f4 100644
--- a/strix/agents/prompts/system_prompt.jinja
+++ b/strix/agents/prompts/system_prompt.jinja
@@ -148,9 +148,7 @@ OPERATIONAL PRINCIPLES:
- Default to recon first. Unless the next step is obvious from context or the user/system gives specific prioritization instructions, begin by mapping the target well before diving into narrow validation or targeted testing
- Prefer established industry-standard tools already available in the sandbox before writing custom scripts
- Do NOT reinvent the wheel with ad hoc Python or shell code when a suitable existing tool can do the job reliably
-- Use the load_skill tool when you need exact vulnerability-specific, protocol-specific, or tool-specific guidance before acting
-- Prefer loading a relevant skill before guessing payloads, workflows, or tool syntax from memory
-- If a task maps cleanly to one or more available skills, load them early and let them guide your next actions
+- Skills relevant to your task are preloaded into this prompt at scan start; refer back to them when you need vulnerability-, protocol-, or tool-specific guidance
- Use custom Python or shell code when you want to dig deeper, automate custom workflows, batch operations, triage results, build target-specific validation, or do work that existing tools do not cover cleanly
- Chain related weaknesses when needed to demonstrate real impact
- Consider business logic and context in validation
diff --git a/strix/entry.py b/strix/entry.py
index 7cc143bd..bc4e4c93 100644
--- a/strix/entry.py
+++ b/strix/entry.py
@@ -21,6 +21,7 @@ from pathlib import Path
from typing import TYPE_CHECKING, Any
from agents import Runner
+from agents.tracing import add_trace_processor
from strix.agents.factory import build_strix_agent, make_child_factory
from strix.orchestration.bus import AgentMessageBus
@@ -31,6 +32,7 @@ from strix.run_config_factory import (
make_run_config,
)
from strix.sandbox import session_manager
+from strix.telemetry.strix_processor import StrixTracingProcessor
if TYPE_CHECKING:
@@ -156,8 +158,10 @@ async def run_strix_scan(
image: str,
sources_path: Path,
tracer: Any | None = None,
+ bus: AgentMessageBus | None = None,
interactive: bool = False,
max_turns: int = STRIX_DEFAULT_MAX_TURNS,
+ model: str = "anthropic/claude-sonnet-4-6",
cleanup_on_exit: bool = True,
) -> RunResult:
"""Run one Strix scan end-to-end against a freshly-prepared sandbox.
@@ -188,9 +192,25 @@ async def run_strix_scan(
scan_id = f"scan-{uuid.uuid4().hex[:8]}"
logger.info("Starting Strix scan %s", scan_id)
- bus = AgentMessageBus()
+ # Caller may pre-create the bus so it can hold a handle (e.g., the
+ # TUI uses it to route stop / chat-input commands). Otherwise we
+ # own the bus internally for the scan's lifetime.
+ if bus is None:
+ bus = AgentMessageBus()
root_id = uuid.uuid4().hex[:8]
+ # Wire SDK tracing into the scan's run-directory ``events.jsonl``.
+ # ``add_trace_processor`` is idempotent at the provider level — if
+ # the user runs multiple scans in one process they each get their
+ # own processor, all writing to their respective run dirs.
+ if tracer is not None:
+ try:
+ run_dir = tracer.get_run_dir() if hasattr(tracer, "get_run_dir") else None
+ if run_dir is not None:
+ add_trace_processor(StrixTracingProcessor(run_dir))
+ except Exception:
+ logger.exception("Failed to register StrixTracingProcessor")
+
bundle = await session_manager.create_or_reuse(
scan_id,
image=image,
@@ -232,10 +252,12 @@ async def run_strix_scan(
sandbox_token=bundle["bearer"],
tool_server_host_port=bundle["tool_server_host_port"],
caido_host_port=bundle["caido_host_port"],
+ caido_capability=bundle.get("capability"),
agent_id=root_id,
agent_name="strix",
parent_id=None,
tracer=tracer,
+ model=model,
max_turns=max_turns,
is_whitebox=is_whitebox,
diff_scope=diff_scope,
@@ -246,6 +268,7 @@ async def run_strix_scan(
run_config = make_run_config(
sandbox_session=bundle["session"],
sandbox_client=bundle["client"],
+ model=model,
)
task_text = _build_root_task(scan_config)
diff --git a/strix/interface/tool_components/__init__.py b/strix/interface/tool_components/__init__.py
index c8b60073..cb8aeeac 100644
--- a/strix/interface/tool_components/__init__.py
+++ b/strix/interface/tool_components/__init__.py
@@ -4,7 +4,6 @@ from . import (
browser_renderer,
file_edit_renderer,
finish_renderer,
- load_skill_renderer,
notes_renderer,
proxy_renderer,
python_renderer,
@@ -29,7 +28,6 @@ __all__ = [
"file_edit_renderer",
"finish_renderer",
"get_tool_renderer",
- "load_skill_renderer",
"notes_renderer",
"proxy_renderer",
"python_renderer",
diff --git a/strix/interface/tool_components/load_skill_renderer.py b/strix/interface/tool_components/load_skill_renderer.py
deleted file mode 100644
index 41a18682..00000000
--- a/strix/interface/tool_components/load_skill_renderer.py
+++ /dev/null
@@ -1,33 +0,0 @@
-from typing import Any, ClassVar
-
-from rich.text import Text
-from textual.widgets import Static
-
-from .base_renderer import BaseToolRenderer
-from .registry import register_tool_renderer
-
-
-@register_tool_renderer
-class LoadSkillRenderer(BaseToolRenderer):
- tool_name: ClassVar[str] = "load_skill"
- css_classes: ClassVar[list[str]] = ["tool-call", "load-skill-tool"]
-
- @classmethod
- def render(cls, tool_data: dict[str, Any]) -> Static:
- args = tool_data.get("args", {})
- status = tool_data.get("status", "completed")
-
- requested = args.get("skills", "")
-
- text = Text()
- text.append("◇ ", style="#10b981")
- text.append("loading skill", style="dim")
-
- if requested:
- text.append(" ")
- text.append(requested, style="#10b981")
- elif not tool_data.get("result"):
- text.append("\n ")
- text.append("Loading...", style="dim")
-
- return Static(text, classes=cls.get_css_classes(status))
diff --git a/strix/interface/tui.py b/strix/interface/tui.py
index 3db4ba11..179f4c5e 100644
--- a/strix/interface/tui.py
+++ b/strix/interface/tui.py
@@ -711,6 +711,13 @@ class StrixTUIApp(App): # type: ignore[misc]
self.tracer.set_scan_config(self.scan_config)
set_global_tracer(self.tracer)
+ # Pre-create the bus here (rather than letting ``run_strix_scan``
+ # build its own) so the TUI can hold a handle for stop / chat
+ # routing while the scan loop runs in a worker thread.
+ from strix.orchestration.bus import AgentMessageBus
+
+ self.bus: AgentMessageBus = AgentMessageBus()
+
self.agent_nodes: dict[str, TreeNode] = {}
self._displayed_agents: set[str] = set()
@@ -720,6 +727,7 @@ class StrixTUIApp(App): # type: ignore[misc]
self._last_streaming_len: dict[str, int] = {}
self._scan_thread: threading.Thread | None = None
+ self._scan_loop: asyncio.AbstractEventLoop | None = None
self._scan_stop_event = threading.Event()
self._scan_completed = threading.Event()
@@ -1496,6 +1504,10 @@ class StrixTUIApp(App): # type: ignore[misc]
try:
loop = asyncio.new_event_loop()
asyncio.set_event_loop(loop)
+ # Stash the loop so synchronous TUI handlers (stop /
+ # chat) can submit bus coroutines onto it from the
+ # main thread.
+ self._scan_loop = loop
try:
if not self._scan_stop_event.is_set():
@@ -1508,6 +1520,7 @@ class StrixTUIApp(App): # type: ignore[misc]
image=str(image),
sources_path=sources_path,
tracer=self.tracer,
+ bus=self.bus,
interactive=True,
),
)
@@ -1827,10 +1840,6 @@ class StrixTUIApp(App): # type: ignore[misc]
metadata={"interrupted": True},
)
- # TODO: route user→agent messages through the AgentMessageBus
- # once the TUI has a handle on it. The bus currently lives
- # inside ``run_strix_scan`` scope only.
-
if self.tracer:
self.tracer.log_chat_message(
content=message,
@@ -1838,11 +1847,18 @@ class StrixTUIApp(App): # type: ignore[misc]
agent_id=self.selected_agent_id,
)
- logging.warning(
- "User-message-to-agent dispatch is not wired post-migration; "
- "message %r logged to tracer but not delivered.",
- message,
- )
+ # Route to the agent's bus inbox. The scan loop runs on a
+ # worker thread; ``run_coroutine_threadsafe`` submits the
+ # coroutine onto that loop and returns immediately so the TUI
+ # stays responsive.
+ if self._scan_loop is not None and not self._scan_loop.is_closed():
+ asyncio.run_coroutine_threadsafe(
+ self.bus.send(
+ self.selected_agent_id,
+ {"from": "user", "content": message, "type": "instruction"},
+ ),
+ self._scan_loop,
+ )
self._displayed_events.clear()
self._update_chat_view()
@@ -1941,11 +1957,12 @@ class StrixTUIApp(App): # type: ignore[misc]
return agent_name, False
def action_confirm_stop_agent(self, agent_id: str) -> None:
- # TODO: route to ``bus.cancel_descendants(agent_id)`` once the TUI
- # has a handle on the AgentMessageBus.
- logging.warning(
- "Stop-agent dispatch is not wired post-migration; agent %s left running.",
- agent_id,
+ if self._scan_loop is None or self._scan_loop.is_closed():
+ logging.warning("No active scan loop; cannot stop agent %s", agent_id)
+ return
+ asyncio.run_coroutine_threadsafe(
+ self.bus.cancel_descendants(agent_id),
+ self._scan_loop,
)
def action_custom_quit(self) -> None:
diff --git a/strix/orchestration/hooks.py b/strix/orchestration/hooks.py
index 6fefe700..919ccdfb 100644
--- a/strix/orchestration/hooks.py
+++ b/strix/orchestration/hooks.py
@@ -84,14 +84,31 @@ class StrixOrchestrationHooks(RunHooks[Any]):
agent: Any,
response: ModelResponse,
) -> None:
+ del agent
try:
ctx = context.context
if not isinstance(ctx, dict):
return
- bus = ctx.get("bus")
+ usage = getattr(response, "usage", None)
agent_id = ctx.get("agent_id")
+ bus = ctx.get("bus")
if bus is not None and agent_id is not None:
- await bus.record_usage(agent_id, getattr(response, "usage", None))
+ await bus.record_usage(agent_id, usage)
+ tracer = ctx.get("tracer")
+ if tracer is not None and usage is not None and hasattr(tracer, "record_llm_usage"):
+ cached = 0
+ details = getattr(usage, "input_tokens_details", None)
+ if details is not None:
+ cached = int(getattr(details, "cached_tokens", 0) or 0)
+ tracer.record_llm_usage(
+ agent_id=str(agent_id or "unknown"),
+ input_tokens=int(getattr(usage, "input_tokens", 0) or 0),
+ output_tokens=int(getattr(usage, "output_tokens", 0) or 0),
+ cached_tokens=cached,
+ cost=0.0, # litellm cost computation lives in the legacy LLM
+ requests=1,
+ bucket="live",
+ )
ctx["turn_count"] = int(ctx.get("turn_count", 0)) + 1
except Exception:
logger.exception("on_llm_end failed")
@@ -101,17 +118,19 @@ class StrixOrchestrationHooks(RunHooks[Any]):
context: AgentHookContext[Any],
agent: Any,
) -> None:
+ # The CaidoCapability is bound to the sandbox session, not the
+ # Agent (we use plain ``Agent``, not ``SandboxAgent``). We stash
+ # it in the context dict at scan-bring-up time so the hook can
+ # await its healthcheck before the first LLM call.
+ del agent
try:
- cap = next(
- (
- c
- for c in (getattr(agent, "capabilities", None) or [])
- if hasattr(c, "_healthcheck_task")
- ),
- None,
- )
- if cap is not None and getattr(cap, "_healthcheck_task", None) is not None:
- await cap._healthcheck_task
+ ctx = context.context
+ if not isinstance(ctx, dict):
+ return
+ cap = ctx.get("caido_capability")
+ task = getattr(cap, "_healthcheck_task", None)
+ if task is not None:
+ await task
except Exception:
logger.exception("on_agent_start failed")
diff --git a/strix/run_config_factory.py b/strix/run_config_factory.py
index cd09f474..8f051afb 100644
--- a/strix/run_config_factory.py
+++ b/strix/run_config_factory.py
@@ -171,6 +171,7 @@ def make_agent_context(
run_id: str | None = None,
sandbox_client: Any | None = None,
agent_factory: Any | None = None,
+ caido_capability: Any | None = None,
) -> dict[str, Any]:
"""Build the per-agent ``context`` dict passed to ``Runner.run(context=...)``.
@@ -190,6 +191,7 @@ def make_agent_context(
"sandbox_token": sandbox_token,
"tool_server_host_port": tool_server_host_port,
"caido_host_port": caido_host_port,
+ "caido_capability": caido_capability,
"agent_id": agent_id,
"agent_name": agent_name,
"parent_id": parent_id,
diff --git a/strix/tools/__init__.py b/strix/tools/__init__.py
index 87ad8d4d..41be98a2 100644
--- a/strix/tools/__init__.py
+++ b/strix/tools/__init__.py
@@ -15,7 +15,6 @@ from .agents_graph import * # noqa: F403
from .browser import * # noqa: F403
from .file_edit import * # noqa: F403
from .finish import * # noqa: F403
-from .load_skill import * # noqa: F403
from .notes import * # noqa: F403
from .proxy import * # noqa: F403
from .python import * # noqa: F403
diff --git a/strix/tools/_state_adapter.py b/strix/tools/_state_adapter.py
index b126c898..c395528e 100644
--- a/strix/tools/_state_adapter.py
+++ b/strix/tools/_state_adapter.py
@@ -8,7 +8,6 @@ adapter object from the run context.
Used by:
- ``tools/todo/tools.py``
- - ``tools/load_skill/tool.py``
- ``tools/finish/tool.py``
"""
diff --git a/strix/tools/agents_graph/tools.py b/strix/tools/agents_graph/tools.py
index bfb0311c..5b493046 100644
--- a/strix/tools/agents_graph/tools.py
+++ b/strix/tools/agents_graph/tools.py
@@ -357,6 +357,7 @@ async def create_agent(
sandbox_token=inner.get("sandbox_token"),
tool_server_host_port=inner.get("tool_server_host_port"),
caido_host_port=inner.get("caido_host_port"),
+ caido_capability=inner.get("caido_capability"),
agent_id=child_id,
agent_name=name,
parent_id=parent_id,
diff --git a/strix/tools/load_skill/__init__.py b/strix/tools/load_skill/__init__.py
deleted file mode 100644
index f220d247..00000000
--- a/strix/tools/load_skill/__init__.py
+++ /dev/null
@@ -1,4 +0,0 @@
-from .load_skill_actions import load_skill
-
-
-__all__ = ["load_skill"]
diff --git a/strix/tools/load_skill/load_skill_actions.py b/strix/tools/load_skill/load_skill_actions.py
deleted file mode 100644
index 54b97175..00000000
--- a/strix/tools/load_skill/load_skill_actions.py
+++ /dev/null
@@ -1,58 +0,0 @@
-from typing import Any
-
-from strix.tools.registry import register_tool
-
-
-@register_tool(sandbox_execution=False)
-def load_skill(agent_state: Any, skills: str) -> dict[str, Any]:
- try:
- from strix.skills import parse_skill_list, validate_requested_skills
-
- requested_skills = parse_skill_list(skills)
- if not requested_skills:
- return {
- "success": False,
- "error": "No skills provided. Pass one or more comma-separated skill names.",
- "requested_skills": [],
- }
-
- validation_error = validate_requested_skills(requested_skills)
- if validation_error:
- return {
- "success": False,
- "error": validation_error,
- "requested_skills": requested_skills,
- "loaded_skills": [],
- }
-
- # Runtime skill injection used to reach into the legacy
- # ``_agent_instances`` registry to mutate the running LLM's
- # active-skills list. The SDK harness owns the agent through
- # ``Runner.run`` and there's no equivalent reach-in API yet —
- # the model still gets a structured success response so it can
- # observe which skills it asked for, even if reload-on-the-fly
- # is a Phase 6 follow-up.
- newly_loaded = list(requested_skills)
- already_loaded: list[str] = []
-
- except Exception as e: # noqa: BLE001
- fallback_requested_skills = (
- requested_skills
- if "requested_skills" in locals()
- else [s.strip() for s in skills.split(",") if s.strip()]
- )
- return {
- "success": False,
- "error": f"Failed to load skill(s): {e!s}",
- "requested_skills": fallback_requested_skills,
- "loaded_skills": [],
- }
- else:
- return {
- "success": True,
- "requested_skills": requested_skills,
- "loaded_skills": requested_skills,
- "newly_loaded_skills": newly_loaded,
- "already_loaded_skills": already_loaded,
- "message": "Skills loaded into this agent prompt context.",
- }
diff --git a/strix/tools/load_skill/load_skill_actions_schema.xml b/strix/tools/load_skill/load_skill_actions_schema.xml
deleted file mode 100644
index aba1aee3..00000000
--- a/strix/tools/load_skill/load_skill_actions_schema.xml
+++ /dev/null
@@ -1,33 +0,0 @@
-<tools>
- <tool name="load_skill">
- <description>Dynamically load one or more skills into the current agent at runtime.
-
-Use this when you need exact guidance right before acting (tool syntax, exploit workflow, or protocol details).
-This updates the current agent's prompt context immediately.</description>
- <details>Accepts one skill or a comma-separated skill bundle. Works for root agents and subagents.
-Examples:
-- Single skill: `xss`
-- Bundle: `sql_injection,business_logic`</details>
- <parameters>
- <parameter name="skills" type="string" required="true">
- <description>Comma-separated list of skills to use for the agent (MAXIMUM 5 skills allowed). Most agents should have at least one skill in order to be useful. Agents should be highly specialized - use 1-3 related skills; up to 5 for complex contexts. {{DYNAMIC_SKILLS_DESCRIPTION}}</description>
- </parameter>
- </parameters>
- <returns type="Dict[str, Any]">
- <description>Response containing: - success: Whether runtime loading succeeded - requested_skills: Skills requested - loaded_skills: Skills validated and applied - newly_loaded_skills: Skills newly injected into prompt - already_loaded_skills: Skills already present in prompt context</description>
- </returns>
- <examples>
- <function=load_skill>
- <parameter=skills>xss</parameter>
- </function>
-
- <function=load_skill>
- <parameter=skills>sql_injection,business_logic</parameter>
- </function>
-
- <function=load_skill>
- <parameter=skills>nmap,httpx</parameter>
- </function>
- </examples>
- </tool>
-</tools>
diff --git a/strix/tools/load_skill/tool.py b/strix/tools/load_skill/tool.py
deleted file mode 100644
index 4c5d051b..00000000
--- a/strix/tools/load_skill/tool.py
+++ /dev/null
@@ -1,46 +0,0 @@
-"""SDK function-tool wrapper for the legacy ``load_skill`` tool.
-
-The legacy implementation reaches into ``_agent_instances`` (a global
-dict the legacy multi-agent orchestrator maintains) to find the running
-``Agent`` instance and call ``agent.llm.add_skills(...)``. That global
-goes away under the SDK migration — Phase 3 will replace it with a
-context-keyed registry, and this wrapper will be updated to read from
-that registry.
-
-For Phase 2 we ship the wrapper as-is. The legacy function falls back
-to a clean error path when the agent instance lookup fails, so the
-tool degrades gracefully ("Could not find running agent instance...")
-until Phase 3 lands. That's better than crashing or stubbing out the
-tool entirely — the model still gets a structured error it can react to.
-"""
-
-from __future__ import annotations
-
-import asyncio
-import json
-from typing import Any
-
-from agents import RunContextWrapper
-
-from strix.tools._decorator import strix_tool
-from strix.tools._state_adapter import adapter_from_ctx
-from strix.tools.load_skill import load_skill_actions as _impl
-
-
-def _dump(result: dict[str, Any]) -> str:
- return json.dumps(result, ensure_ascii=False, default=str)
-
-
-@strix_tool(timeout=60)
-async def load_skill(ctx: RunContextWrapper, skills: str) -> str:
- """Load one or more named skills into this agent's prompt context.
-
- Args:
- skills: Comma-separated skill names (max 5). E.g.
- ``"recon,xss,sqli"``. Skill discovery uses
- ``strix.skills.parse_skill_list``.
- """
- state = adapter_from_ctx(ctx)
- return _dump(
- await asyncio.to_thread(_impl.load_skill, agent_state=state, skills=skills),
- )
diff --git a/tests/agents/test_factory.py b/tests/agents/test_factory.py
index 009d9d49..19029805 100644
--- a/tests/agents/test_factory.py
+++ b/tests/agents/test_factory.py
@@ -115,7 +115,6 @@ def test_agent_includes_graph_and_sandbox_tools() -> None:
"web_search",
"str_replace_editor",
"create_vulnerability_report",
- "load_skill",
"browser_action",
"terminal_execute",
"python_action",
diff --git a/tests/tools/test_remaining_local_tools.py b/tests/tools/test_remaining_local_tools.py
index 229e6fb2..3d323829 100644
--- a/tests/tools/test_remaining_local_tools.py
+++ b/tests/tools/test_remaining_local_tools.py
@@ -1,12 +1,7 @@
-"""Phase 2.4 smoke tests for the remaining local SDK tool wrappers.
+"""Smoke tests for the remaining local SDK tool wrappers.
Covers: web_search, file_edit (str_replace_editor + list_files +
-search_files), reporting (create_vulnerability_report), load_skill,
-finish_scan.
-
-Pattern matches test_sdk_local_tools.py — confirm registration as
-``FunctionTool``, exercise the legacy delegation path, verify that
-sandbox-bound tools route through ``post_to_sandbox``.
+search_files), reporting (create_vulnerability_report), finish_scan.
"""
from __future__ import annotations
@@ -25,7 +20,6 @@ from strix.tools.file_edit.tools import (
str_replace_editor,
)
from strix.tools.finish.tool import finish_scan
-from strix.tools.load_skill.tool import load_skill
from strix.tools.reporting.tool import create_vulnerability_report
from strix.tools.web_search.tool import web_search
@@ -69,7 +63,6 @@ def test_all_remaining_tools_are_function_tools() -> None:
list_files,
search_files,
create_vulnerability_report,
- load_skill,
finish_scan,
):
assert isinstance(tool, FunctionTool)
@@ -240,38 +233,6 @@ async def test_create_vulnerability_report_delegates_to_impl() -> None:
assert kwargs["method"] is None
-# --- load_skill ----------------------------------------------------------
-
-
-async def test_load_skill_passes_adapter_with_agent_id() -> None:
- """The wrapper must build the legacy-shaped adapter from ctx.context."""
- captured: dict[str, Any] = {}
-
- def fake_legacy(*, agent_state: Any, skills: str) -> dict[str, Any]:
- captured["agent_id"] = agent_state.agent_id
- captured["skills"] = skills
- return {"success": True, "loaded_skills": ["recon"]}
-
- with patch(
- "strix.tools.load_skill.tool._impl.load_skill",
- side_effect=fake_legacy,
- ):
- out = await _invoke(load_skill, _ctx_for("agent-XYZ"), skills="recon")
-
- assert out["success"] is True
- assert captured["agent_id"] == "agent-XYZ"
- assert captured["skills"] == "recon"
-
-
-async def test_load_skill_with_empty_input() -> None:
- """End-to-end: empty skills string yields the legacy validation error."""
- out = await _invoke(load_skill, _ctx_for(), skills="")
- assert out["success"] is False
- assert "No skills" in out["error"]
-
-
# --- finish_scan ---------------------------------------------------------
diff --git a/tests/tools/test_tool_registration_modes.py b/tests/tools/test_tool_registration_modes.py
index 6812e8eb..ac2346af 100644
--- a/tests/tools/test_tool_registration_modes.py
+++ b/tests/tools/test_tool_registration_modes.py
@@ -58,45 +58,5 @@ def test_sandbox_registers_sandbox_tools_but_not_non_sandbox_tools(
assert "list_requests" in names
assert "create_agent" not in names
assert "finish_scan" not in names
- assert "load_skill" not in names
assert "browser_action" not in names
assert "web_search" not in names
-
-
-def test_load_skill_import_does_not_register_create_agent_in_sandbox(
- monkeypatch: Any,
-) -> None:
- monkeypatch.setenv("STRIX_SANDBOX_MODE", "true")
- monkeypatch.setenv("STRIX_DISABLE_BROWSER", "true")
- monkeypatch.delenv("PERPLEXITY_API_KEY", raising=False)
- monkeypatch.setattr(Config, "load", classmethod(_empty_config_load))
-
- clear_registry()
- for name in list(sys.modules):
- if name == "strix.tools" or name.startswith("strix.tools."):
- sys.modules.pop(name, None)
-
- load_skill_module = importlib.import_module("strix.tools.load_skill.load_skill_actions")
- registry = importlib.import_module("strix.tools.registry")
-
- names_before = set(registry.get_tool_names())
- assert "load_skill" not in names_before
- assert "create_agent" not in names_before
-
- state_type = type(
- "DummyState",
- (),
- {
- "agent_id": "agent_test",
- "context": {},
- "update_context": lambda self, key, value: self.context.__setitem__(key, value),
- },
- )
- result = load_skill_module.load_skill(state_type(), "nmap")
-
- names_after = set(registry.get_tool_names())
- assert "create_agent" not in names_after
- # load_skill no longer reaches into the legacy _agent_instances global —
- # it returns ``success=True`` with the requested skills echoed back.
- assert result["success"] is True
- assert result["loaded_skills"] == ["nmap"]
diff --git a/uv.lock b/uv.lock
index d6c3487a..59e20786 100644
--- a/uv.lock
+++ b/uv.lock
@@ -5441,10 +5441,8 @@ dependencies = [
{ name = "requests" },
{ name = "rich" },
{ name = "scrubadub" },
- { name = "tenacity" },
{ name = "textual" },
{ name = "traceloop-sdk" },
- { name = "xmltodict" },
]
[package.optional-dependencies]
@@ -5501,11 +5499,9 @@ requires-dist = [
{ name = "requests", specifier = ">=2.32.0" },
{ name = "rich" },
{ name = "scrubadub", specifier = ">=2.0.1" },
- { name = "tenacity", specifier = ">=9.0.0" },
{ name = "textual", specifier = ">=6.0.0" },
{ name = "traceloop-sdk", specifier = ">=0.53.0" },
{ name = "uvicorn", marker = "extra == 'sandbox'" },
- { name = "xmltodict", specifier = ">=0.13.0" },
]
provides-extras = ["vertex", "sandbox"]
@@ -6081,15 +6077,6 @@ wheels = [
{ url = "https://files.pythonhosted.org/packages/3a/0c/3662f4a66880196a590b202f0db82d919dd2f89e99a27fadef91c4a33d41/xlsxwriter-3.2.9-py3-none-any.whl", hash = "sha256:9a5db42bc5dff014806c58a20b9eae7322a134abb6fce3c92c181bfb275ec5b3", size = 175315, upload-time = "2025-09-16T00:16:20.108Z" },
]
-[[package]]
-name = "xmltodict"
-version = "1.0.4"
-source = { registry = "https://pypi.org/simple" }
-sdist = { url = "https://files.pythonhosted.org/packages/19/70/80f3b7c10d2630aa66414bf23d210386700aa390547278c789afa994fd7e/xmltodict-1.0.4.tar.gz", hash = "sha256:6d94c9f834dd9e44514162799d344d815a3a4faec913717a9ecbfa5be1bb8e61", size = 26124, upload-time = "2026-02-22T02:21:22.074Z" }
-wheels = [
- { url = "https://files.pythonhosted.org/packages/38/34/98a2f52245f4d47be93b580dae5f9861ef58977d73a79eb47c58f1ad1f3a/xmltodict-1.0.4-py3-none-any.whl", hash = "sha256:a4a00d300b0e1c59fc2bfccb53d7b2e88c32f200df138a0dd2229f842497026a", size = 13580, upload-time = "2026-02-22T02:21:21.039Z" },
-]
-
[[package]]
name = "yarl"
version = "1.23.0"