Files
strix/pyproject.toml
T
0xallam d538acf66b feat(orchestration): always-on resume across the agent graph
A scan that crashes or is stopped can now be resumed by re-invoking
``strix`` with the same ``--run-name``. Resume is implicit — presence
of ``{run_dir}/bus.json`` triggers it. To force a fresh start, delete
the run dir.

What survives a process restart with the same scan_id:

  * Root agent's LLM history — already worked (root SDK SQLiteSession).
  * Every non-terminal subagent's LLM history — new. ``create_agent``
    now opens SQLiteSession(session_id=child_id,
    db_path={run_dir}/sessions/{child_id}.db) per child and passes it
    to ``run_with_continuation``.
  * Bus topology — new. ``AgentMessageBus`` gains snapshot/restore/
    _maybe_snapshot async methods plus a ``metadata`` field that holds
    per-agent {task, skills, is_whitebox, scan_mode, diff_scope}.
    ``register``, ``finalize``, ``park``, and ``mark_llm_failed`` each
    call ``_maybe_snapshot`` to atomically persist the bus to
    {run_dir}/bus.json (tempfile + Path.replace).
  * Vulnerability reports — new. ``ScanArtifactWriter._write_
    vulnerabilities`` now also writes ``vulnerabilities.json``
    (atomic). ``Tracer.hydrate_from_run_dir`` reads it on resume so
    new vuln-NNNN ids don't collide with prior on-disk files.

What does not survive: the sandbox container itself (fresh per
process), so ``/workspace/scratch`` and Caido state are lost.
``/workspace/sources`` re-mounts from the host so source code is
unchanged.

``orchestration/scan.py:run_strix_scan`` does the actual resume:
  1. Resolve run_dir up front; if bus.json exists it's a resume.
  2. Acquire {run_dir}/.lock (fcntl.flock) so a second strix process
     can't run concurrently on the same scan_id.
  3. ``bus.set_snapshot_path(...)``, ``tracer.hydrate_from_run_dir()``.
  4. On resume: load + bus.restore, find root_id from snapshot (the
     agent with parent_of[id] is None), spawn the sandbox, skip the
     root's bus.register (already in snapshot).
  5. ``_respawn_subagents`` walks every agent with status in
     running/waiting/llm_failed: reopens its SQLiteSession, rebuilds
     the child agent via the captured factory, builds run config /
     context, asyncio.create_task the run with initial_input=[] so
     the SDK replays from session. Per-child failure (missing/corrupt
     DB, factory raises) finalizes that child as crashed and continues.
  6. Open root SQLiteSession at the same path, run the root with
     initial_input=[] on resume (or the formatted root task on a
     fresh run), and let SDK replay drive the next turn.
  7. ``finally``: close every per-agent session, take a final
     snapshot, tear down sandbox, release the lock.

HARNESS_WIKI.md updated with the new run-dir layout (sessions/,
bus.json, vulnerabilities.json, .lock) and the resume contract.

Net: +500 LoC across 7 files. No new deps.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 00:29:37 -07:00

345 lines
11 KiB
TOML

[project]
name = "strix-agent"
version = "0.8.3"
description = "Open-source AI Hackers for your apps"
readme = "README.md"
license = "Apache-2.0"
requires-python = ">=3.12"
authors = [
{ name = "Strix", email = "hi@usestrix.com" },
]
keywords = [
"cybersecurity",
"security",
"vulnerability",
"scanner",
"pentest",
"agent",
"ai",
"cli",
]
classifiers = [
"Development Status :: 3 - Alpha",
"Intended Audience :: Information Technology",
"Intended Audience :: System Administrators",
"Topic :: Security",
"License :: OSI Approved :: Apache Software License",
"Environment :: Console",
"Programming Language :: Python",
"Programming Language :: Python :: 3",
"Programming Language :: Python :: 3 :: Only",
"Programming Language :: Python :: 3.12",
"Programming Language :: Python :: 3.13",
"Programming Language :: Python :: 3.14",
]
dependencies = [
"openai-agents[litellm]==0.14.6",
"pydantic>=2.11.3",
"pydantic-settings>=2.13.0",
"rich",
"docker>=7.1.0",
"textual>=6.0.0",
"requests>=2.32.0",
"cvss>=3.2",
"caido-sdk-client>=0.2.0",
]
[project.scripts]
strix = "strix.interface.main:main"
[dependency-groups]
dev = [
"mypy>=1.16.0",
"ruff>=0.11.13",
"pyright>=1.1.401",
"bandit>=1.8.3",
"pre-commit>=4.2.0",
"pyinstaller>=6.17.0; python_version >= '3.12' and python_version < '3.15'",
]
[build-system]
requires = ["hatchling"]
build-backend = "hatchling.build"
[tool.hatch.build.targets.wheel]
packages = ["strix"]
# ============================================================================
# Type Checking Configuration
# ============================================================================
[tool.mypy]
python_version = "3.12"
strict = true
strict_optional = true
warn_redundant_casts = true
warn_unused_ignores = true
warn_return_any = true
warn_unreachable = true
disallow_untyped_defs = true
disallow_any_generics = true
disallow_subclassing_any = true
disallow_untyped_calls = true
disallow_incomplete_defs = true
check_untyped_defs = true
disallow_untyped_decorators = true
no_implicit_optional = true
warn_unused_configs = true
show_error_codes = true
show_column_numbers = true
pretty = true
# Allow some flexibility for third-party libraries
[[tool.mypy.overrides]]
module = [
"litellm.*",
"rich.*",
"jinja2.*",
"textual.*",
"cvss.*",
"docker.*",
"caido_sdk_client.*",
"pydantic_settings.*",
]
ignore_missing_imports = true
disable_error_code = ["import-untyped"]
# ============================================================================
# Ruff Configuration (Fast Python Linter & Formatter)
# ============================================================================
[tool.ruff]
target-version = "py312"
line-length = 100
extend-exclude = [
".git",
".mypy_cache",
".ruff_cache",
"__pycache__",
"build",
"dist",
"migrations",
]
[tool.ruff.lint]
# Enable comprehensive rule sets
select = [
"E", # pycodestyle errors
"W", # pycodestyle warnings
"F", # Pyflakes
"I", # isort
"N", # pep8-naming
"UP", # pyupgrade
"YTT", # flake8-2020
"S", # flake8-bandit
"BLE", # flake8-blind-except
"FBT", # flake8-boolean-trap
"B", # flake8-bugbear
"A", # flake8-builtins
"COM", # flake8-commas
"C4", # flake8-comprehensions
"DTZ", # flake8-datetimez
"T10", # flake8-debugger
"EM", # flake8-errmsg
"FA", # flake8-future-annotations
"ISC", # flake8-implicit-str-concat
"ICN", # flake8-import-conventions
"G", # flake8-logging-format
"INP", # flake8-no-pep420
"PIE", # flake8-pie
"T20", # flake8-print
"PYI", # flake8-pyi
"Q", # flake8-quotes
"RSE", # flake8-raise
"RET", # flake8-return
"SLF", # flake8-self
"SIM", # flake8-simplify
"TID", # flake8-tidy-imports
"TCH", # flake8-type-checking
"ARG", # flake8-unused-arguments
"PTH", # flake8-use-pathlib
"ERA", # eradicate
"PD", # pandas-vet
"PGH", # pygrep-hooks
"PL", # Pylint
"TRY", # tryceratops
"FLY", # flynt
"PERF", # Perflint
"RUF", # Ruff-specific rules
]
ignore = [
"S101", # Use of assert
"S104", # Possible binding to all interfaces
"S301", # Use of pickle
"COM812", # Missing trailing comma (handled by formatter)
"ISC001", # Single line implicit string concatenation (handled by formatter)
"PLR0913", # Too many arguments to function call
"TRY003", # Avoid specifying long messages outside the exception class
"EM101", # Exception must not use a string literal
"EM102", # Exception must not use an f-string literal
"FBT001", # Boolean positional arg in function definition
"FBT002", # Boolean default positional argument in function definition
"G004", # Logging statement uses f-string
"PLR2004", # Magic value used in comparison
"SLF001", # Private member accessed
]
[tool.ruff.lint.per-file-ignores]
# Lazy imports inside functions to avoid circular dependency with
# strix.telemetry / strix.llm.dedupe / cvss.
"strix/tools/notes/tools.py" = ["PLC0415", "TC002"]
"strix/tools/finish/tool.py" = ["PLC0415", "TC002"]
"strix/tools/reporting/tool.py" = ["PLC0415", "TC002"]
"strix/tools/**/*.py" = [
"ARG001", # Unused function argument (tools may have unused args for interface consistency)
]
# SDK lifecycle hooks have a fixed signature — unused args are part of the contract.
"strix/orchestration/hooks.py" = [
"ARG002", # Unused method argument (RunHooks signature is set by SDK)
"TC002", # ModelResponse, RunContextWrapper, AgentHookContext are annotation-only
]
# Anthropic cache wrapper inherits from LitellmModel; signature shadows builtin `input`.
"strix/llm/anthropic_cache_wrapper.py" = [
"A002", # Argument shadows builtin (parent signature uses `input`)
"TC002", # Many SDK types are imports-for-annotations only
"TC003", # collections.abc.AsyncIterator imported for return type
]
# Custom Docker subclass duplicates parent body; some imports are for annotations.
# Backend factories import their backend's deps lazily so deployments
# that pick a different backend don't need every backend's libs installed.
"strix/runtime/backends.py" = ["PLC0415"]
# The vulnerability MD renderer is a long monolithic format function;
# splitting per-section would obscure the structure without simplifying.
"strix/telemetry/scan_artifacts.py" = ["PLR0912", "PLR0915", "PERF401"]
"strix/runtime/docker_client.py" = [
"TC002", # Manifest, Container imported for annotations
"TC003", # uuid imported for annotation
]
"strix/llm/multi_provider_setup.py" = [
"TC002", # Model, ModelProvider imported for annotations
]
# SDK function-tool wrappers: the SDK calls get_type_hints() at registration
# time to derive the JSON schema, which evaluates annotations at runtime —
# so RunContextWrapper / Tool / TResponseInputItem must be imported eagerly,
# not under TYPE_CHECKING.
"strix/tools/todo/tools.py" = ["TC002"]
"strix/tools/thinking/tool.py" = ["TC002"]
"strix/tools/web_search/tool.py" = ["TC002"]
"strix/tools/proxy/tools.py" = ["TC002", "PLR0911"]
# Driver script writes to /tmp inside the sandbox container — single-user,
# isolated rootfs, so the multi-user race S108 warns about doesn't apply.
"strix/tools/python/tool.py" = ["S108", "TC002"]
"strix/tools/agents_graph/tools.py" = ["TC002"]
"strix/agents/factory.py" = ["TC002"]
# Entry point: ``Path`` is used at runtime by the typing of the
# session_manager call; importing under TYPE_CHECKING would defer
# resolution past where mypy needs it. ``_build_root_task`` legitimately
# walks every supported target type — splitting it into per-type
# helpers would add indirection without simplifying anything.
"strix/orchestration/scan.py" = ["TC003", "PLR0912", "PLR0915", "PLC0415"]
# Tracer carries a long event surface and a runtime ``Callable``
# annotation on ``vulnerability_found_callback``.
"strix/telemetry/tracer.py" = ["TC003", "PLR0912", "PLR0915", "E501"]
# Interface utility branches per scope-mode / target-type combination;
# splitting would obscure the decision tree without simplifying it.
"strix/interface/utils.py" = ["PLR0912", "BLE001", "PLC0415"]
# CLI / TUI / main keep extensive lazy imports + broad exception
# swallows for resilience around terminal-rendering errors.
"strix/interface/cli.py" = ["BLE001", "PLC0415"]
"strix/interface/tui.py" = ["BLE001", "PLC0415", "PLR0912", "PLR0915", "SIM105"]
"strix/interface/main.py" = ["BLE001", "PLC0415", "PLR0912", "PLR0915"]
"strix/interface/streaming_parser.py" = ["PLC0415"]
"strix/interface/tool_components/agent_message_renderer.py" = ["PLC0415"]
[tool.ruff.lint.isort]
force-single-line = false
lines-after-imports = 2
known-first-party = ["strix"]
known-third-party = ["pydantic"]
[tool.ruff.lint.pylint]
max-args = 8
[tool.ruff.format]
quote-style = "double"
indent-style = "space"
skip-magic-trailing-comma = false
line-ending = "auto"
# ============================================================================
# PyRight Configuration (Alternative type checker)
# ============================================================================
[tool.pyright]
include = ["strix"]
exclude = ["**/__pycache__", "build", "dist"]
pythonVersion = "3.12"
pythonPlatform = "Linux"
typeCheckingMode = "strict"
reportMissingImports = true
reportMissingTypeStubs = false
reportGeneralTypeIssues = true
reportPropertyTypeMismatch = true
reportFunctionMemberAccess = true
reportMissingParameterType = true
reportMissingTypeArgument = true
reportIncompatibleMethodOverride = true
reportIncompatibleVariableOverride = true
reportInconsistentConstructor = true
reportOverlappingOverload = true
reportConstantRedefinition = true
reportImportCycles = true
reportUnusedImport = true
reportUnusedClass = true
reportUnusedFunction = true
reportUnusedVariable = true
reportDuplicateImport = true
# ============================================================================
# Black Configuration (Code Formatter)
# ============================================================================
[tool.black]
line-length = 100
target-version = ['py312']
include = '\\.pyi?$'
extend-exclude = '''
/(
# directories
\.eggs
| \.git
| \.hg
| \.mypy_cache
| \.tox
| \.venv
| build
| dist
)/
'''
# ============================================================================
# isort Configuration (Import Sorting)
# ============================================================================
[tool.isort]
profile = "black"
line_length = 100
multi_line_output = 3
include_trailing_comma = true
force_grid_wrap = 0
use_parentheses = true
ensure_newline_before_comments = true
known_first_party = ["strix"]
known_third_party = ["pydantic", "litellm"]
# ============================================================================
# Bandit Configuration (Security Linting)
# ============================================================================
[tool.bandit]
exclude_dirs = ["docs", "build", "dist"]
skips = ["B101", "B601", "B404", "B603", "B607"] # Skip assert, shell injection, subprocess import and partial path checks
severity = "medium"