deer-flow/backend/tests/test_local_sandbox_encoding.py
Xinmin Zeng 09988caf95
feat(skills): request-scoped secrets for skills (closes #3861) (#3871)
* feat(sandbox): per-call env injection + platform-secret scrubbing for skills

Add an env parameter to Sandbox.execute_command (abstract + local + AIO) so request-scoped secrets can be injected into skill subprocesses, and scrub platform credentials (*KEY*/*SECRET*/*TOKEN*/*PASSWORD*/*CREDENTIAL*) from the inherited environment by default so scoped injection is not security theatre. LocalSandbox always passes an explicit scrubbed env; AioSandbox routes env-bearing commands through bash.exec(env=) on a fresh session and leaves the legacy persistent-shell path unchanged. Part of #3861. BEHAVIOR CHANGE: execute_command no longer inherits the full os.environ; Windows encoding tests updated to assert the scrubbed dict.

* feat(skills): parse required-secrets frontmatter declaration

Add SecretRequirement and Skill.required_secrets, and parse the required-secrets SKILL.md frontmatter field (a string list or {name, optional} mappings), dropping malformed entries with a warning so one bad declaration does not invalidate the skill. The declared name is both the context.secrets key and the env var injected at activation. Part of #3861.

* feat(runtime): request-scoped secret carrier (context.secrets)

Add SECRETS_CONTEXT_KEY + extract_request_secrets, centralising the context.secrets carrier contract. The existing context passthrough (build_run_config -> _build_runtime_context) already carries the sub-key to runtime.context without mirroring it into configurable; characterization tests lock that behaviour. Part of #3861.

* feat(skills): inject declared secrets at slash-activation into bash env

Binding point A: when a skill is slash-activated, SkillActivationMiddleware resolves its declared required-secrets against the request's context.secrets and writes the per-run injection set to runtime.context. The bash tool forwards that set to execute_command(env=). A skill cannot harvest a host platform credential (is_host_platform_secret guard, cf. GHSA-rhgp-j443-p4rf), and injected values are redacted from bash output (mask_secret_values) so an echoed secret never re-enters the prompt/trace. Part of #3861.

* test(skills): lock the five secret leak surfaces + add trace redaction helper

Regression tests assert the secret value is absent from all five surfaces: prompt (activation message), checkpoint (graph state vs context separation), audit (journal records names only), trace (metadata builder never copies context; never mirrored to configurable), and stdout (mask_secret_values). Add redact_secret_context_keys as a defensive helper for any context serialization. Part of #3861.

* docs(backend): document request-scoped secrets for skills

Add Request-Scoped Secrets subsection (Skills) + env policy note (Sandbox) and the execute_command(env=) signature change, per the doc-sync policy. Part of #3861.

* fix(skills): close gaps found by end-to-end verification of request-scoped secrets

Real-gateway e2e + independent review of #3861 surfaced three defects, now fixed:

1. Slash activation never fired in the live chain. InputSanitizationMiddleware
   wraps user input in BEGIN/END markers before SkillActivationMiddleware sees it,
   and the original text was only preserved when an upload or IM channel set it.
   For a plain text message the slash command became undetectable, so no secret
   was ever resolved. Fix: the sanitizer now setdefaults the pre-wrap text into
   ORIGINAL_USER_CONTENT_KEY (additive; sanitization behaviour unchanged), so
   slash activation works for all messages. Pre-existing latent bug surfaced here.

2. The raw request config (with context.secrets) was persisted to runs.kwargs_json
   and echoed by the run API (RunResponse.kwargs). Fix: redact_config_secrets()
   strips secret-bearing context keys from the persisted/echoed copy in start_run;
   the live config that drives the run keeps them. build_run_config now also sets
   configurable.thread_id on the context path (the checkpointer requires it).

3. Connection-string credentials (DATABASE_URL, REDIS_URL, SENTRY_DSN, GH_PAT, ...)
   were not scrubbed from the inherited sandbox env. Fix: env_policy adds a *DSN*
   pattern plus an explicit connection-string denylist (no blanket *URL* — benign
   service URLs stay readable).

Verified end-to-end via a real gateway run (real LLM + skill activation + bash):
the secret reaches the sandbox subprocess and appears in NONE of prompt, trace,
checkpoint, audit, stdout, runs.kwargs_json, or the run API. Part of #3861.

* docs(backend): document the env scrub, persistence redaction, and sanitizer interaction

Sync the Request-Scoped Secrets section with the verification-driven fixes: inherited-env scrub (incl. connection-string denylist), run-record/run-API redaction as the 6th sealed leak surface, and the sanitizer preserving original content so slash activation fires. Part of #3861.

* fix(skills): inject caller secret over scrubbed host value; drop redundant host-name guard

A real-world demo (a skill calling a third-party cloud API with a request-scoped
key) exposed that the is_host_platform_secret guard was both wrong and harmful:
it refused to inject a caller-supplied secret whenever a same-named variable
existed in the Gateway env — which is exactly the #3861 use case (a per-user key
overriding a shared platform key). The guard was also redundant: build_sandbox_env
already scrubs secret-looking names from the inherited env before injection, so a
skill can never read a host credential — it only ever receives the caller's value.

Remove the guard; the injected (caller) value simply wins over the scrubbed host
value. Verified end-to-end: the agent called the real cloud API successfully with
the caller's key, the host's same-named key was scrubbed and never used, and the
caller's key leaked to none of the surfaces. Part of #3861.

* fix(skills): address review on request-scoped secrets (#3861)

Review fixes from PR #3871:

- E2BSandbox.execute_command now accepts env/timeout and routes them to
  commands.run(envs=, timeout=). The bash tool passes env= unconditionally,
  so the prior signature (command only) raised TypeError on every e2b bash
  call and broke e2b deployments entirely. env=None stays backward-compatible.
- SkillActivationMiddleware clears the active-secret set before resolving each
  activation, so a later skill in the same run never inherits an earlier
  skill's injection set (the #3861 contract: a skill only receives what the
  caller supplied AND that skill declared).
- AioSandbox env path uses a dedicated _DEFAULT_HARD_TIMEOUT — bash.exec exposes
  no idle/no-change timeout, so the prior reuse of the legacy idle constant
  conflated wall-clock vs idle semantics. The env path also retries on the
  ErrorObservation signature now, sharing the legacy persistent-shell recovery
  contract.
- mask_secret_values skips values below a minimum length floor so a short
  declared secret (e.g. "42") cannot shred unrelated bytes (exit codes,
  timestamps, sizes) of tool output. The secret is still injected into the
  subprocess; only the output mask skips it.

session_id reuse on the env path is intentionally NOT added: a shared session
could let request-scoped secrets ride the session env into later commands,
which the SDK does not contractually forbid. The fresh-session choice matches
the LocalSandbox model (each call is a fresh subprocess); the trade-off
(consecutive env-bearing calls do not share cwd/venv/exports) is documented on
_execute_with_env.
2026-07-03 07:51:22 +08:00

201 lines
7.5 KiB
Python

import builtins
from types import SimpleNamespace
import deerflow.sandbox.local.local_sandbox as local_sandbox
from deerflow.sandbox.local.local_sandbox import LocalSandbox
def _open(base, file, mode="r", *args, **kwargs):
if "b" in mode:
return base(file, mode, *args, **kwargs)
return base(file, mode, *args, encoding=kwargs.pop("encoding", "gbk"), **kwargs)
def test_read_file_uses_utf8_on_windows_locale(tmp_path, monkeypatch):
path = tmp_path / "utf8.txt"
text = "\u201cutf8\u201d"
path.write_text(text, encoding="utf-8")
base = builtins.open
monkeypatch.setattr(local_sandbox, "open", lambda file, mode="r", *args, **kwargs: _open(base, file, mode, *args, **kwargs), raising=False)
assert LocalSandbox("t").read_file(str(path)) == text
def test_write_file_uses_utf8_on_windows_locale(tmp_path, monkeypatch):
path = tmp_path / "utf8.txt"
text = "emoji \U0001f600"
base = builtins.open
monkeypatch.setattr(local_sandbox, "open", lambda file, mode="r", *args, **kwargs: _open(base, file, mode, *args, **kwargs), raising=False)
LocalSandbox("t").write_file(str(path), text)
assert path.read_text(encoding="utf-8") == text
def test_get_shell_prefers_posix_shell_from_path_before_windows_fallback(monkeypatch):
monkeypatch.setattr(local_sandbox.os, "name", "nt")
monkeypatch.setattr(LocalSandbox, "_find_first_available_shell", lambda candidates: r"C:\Program Files\Git\bin\sh.exe" if candidates == ("/bin/zsh", "/bin/bash", "/bin/sh", "sh") else None)
assert LocalSandbox._get_shell() == r"C:\Program Files\Git\bin\sh.exe"
def test_get_shell_uses_powershell_fallback_on_windows(monkeypatch):
calls: list[tuple[str, ...]] = []
def fake_find(candidates: tuple[str, ...]) -> str | None:
calls.append(candidates)
if candidates == ("/bin/zsh", "/bin/bash", "/bin/sh", "sh"):
return None
return r"C:\Windows\System32\WindowsPowerShell\v1.0\powershell.exe"
monkeypatch.setattr(local_sandbox.os, "name", "nt")
monkeypatch.setattr(local_sandbox.os, "environ", {"SystemRoot": r"C:\Windows"})
monkeypatch.setattr(LocalSandbox, "_find_first_available_shell", fake_find)
assert LocalSandbox._get_shell() == r"C:\Windows\System32\WindowsPowerShell\v1.0\powershell.exe"
assert calls[1] == (
"pwsh",
"pwsh.exe",
"powershell",
"powershell.exe",
r"C:\Windows\System32\WindowsPowerShell\v1.0\powershell.exe",
"cmd.exe",
)
def test_get_shell_uses_cmd_as_last_windows_fallback(monkeypatch):
def fake_find(candidates: tuple[str, ...]) -> str | None:
if candidates == ("/bin/zsh", "/bin/bash", "/bin/sh", "sh"):
return None
return r"C:\Windows\System32\cmd.exe"
monkeypatch.setattr(local_sandbox.os, "name", "nt")
monkeypatch.setattr(local_sandbox.os, "environ", {"SystemRoot": r"C:\Windows"})
monkeypatch.setattr(LocalSandbox, "_find_first_available_shell", fake_find)
assert LocalSandbox._get_shell() == r"C:\Windows\System32\cmd.exe"
def test_execute_command_uses_powershell_command_mode_on_windows(monkeypatch):
calls: list[tuple[object, dict]] = []
def fake_run(*args, **kwargs):
calls.append((args[0], kwargs))
return SimpleNamespace(stdout="ok", stderr="", returncode=0)
monkeypatch.setattr(local_sandbox.os, "name", "nt")
monkeypatch.setattr(local_sandbox.os, "environ", {"PATH": r"C:\Windows", "OPENAI_API_KEY": "should-not-leak"})
monkeypatch.setattr(LocalSandbox, "_get_shell", staticmethod(lambda: r"C:\Windows\System32\WindowsPowerShell\v1.0\powershell.exe"))
monkeypatch.setattr(local_sandbox.subprocess, "run", fake_run)
output = LocalSandbox("t").execute_command("Write-Output hello")
assert output == "ok"
# Platform secrets are scrubbed from the inherited environment even on the
# Windows PowerShell path (#3861); benign PATH is preserved and the env is an
# explicit scrubbed dict, no longer None.
assert calls == [
(
[
r"C:\Windows\System32\WindowsPowerShell\v1.0\powershell.exe",
"-NoProfile",
"-Command",
"Write-Output hello",
],
{
"shell": False,
"capture_output": True,
"text": True,
"timeout": 600,
"env": {"PATH": r"C:\Windows"},
},
)
]
def test_execute_command_uses_posix_shell_command_mode_on_windows(monkeypatch):
calls: list[tuple[object, dict]] = []
def fake_run(*args, **kwargs):
calls.append((args[0], kwargs))
return SimpleNamespace(stdout="ok", stderr="", returncode=0)
monkeypatch.setattr(local_sandbox.os, "name", "nt")
monkeypatch.setattr(local_sandbox.os, "environ", {"PATH": r"C:\Program Files\Git\bin"})
monkeypatch.setattr(LocalSandbox, "_get_shell", staticmethod(lambda: r"C:\Program Files\Git\bin\sh.exe"))
monkeypatch.setattr(local_sandbox.subprocess, "run", fake_run)
output = LocalSandbox("t").execute_command("echo hello")
assert output == "ok"
assert calls == [
(
[r"C:\Program Files\Git\bin\sh.exe", "-c", "echo hello"],
{
"shell": False,
"capture_output": True,
"text": True,
"timeout": 600,
"env": {
"PATH": r"C:\Program Files\Git\bin",
"MSYS_NO_PATHCONV": "1",
"MSYS2_ARG_CONV_EXCL": "*",
},
},
)
]
def test_execute_command_does_not_set_msys_env_for_non_msys_posix_shell_on_windows(monkeypatch):
calls: list[tuple[object, dict]] = []
def fake_run(*args, **kwargs):
calls.append((args[0], kwargs))
return SimpleNamespace(stdout="ok", stderr="", returncode=0)
monkeypatch.setattr(local_sandbox.os, "name", "nt")
monkeypatch.setattr(local_sandbox.os, "environ", {"PATH": r"C:\tools"})
monkeypatch.setattr(LocalSandbox, "_get_shell", staticmethod(lambda: r"C:\tools\busybox\sh.exe"))
monkeypatch.setattr(local_sandbox.subprocess, "run", fake_run)
output = LocalSandbox("t").execute_command("echo /mnt/skills/demo")
assert output == "ok"
# Non-MSYS posix shell adds no MSYS_* vars; the env is the scrubbed inherited
# environment, not None (#3861).
assert calls[0][1]["env"] == {"PATH": r"C:\tools"}
assert "MSYS_NO_PATHCONV" not in calls[0][1]["env"]
def test_execute_command_uses_cmd_command_mode_on_windows(monkeypatch):
calls: list[tuple[object, dict]] = []
def fake_run(*args, **kwargs):
calls.append((args[0], kwargs))
return SimpleNamespace(stdout="ok", stderr="", returncode=0)
monkeypatch.setattr(local_sandbox.os, "name", "nt")
monkeypatch.setattr(local_sandbox.os, "environ", {"PATH": r"C:\Windows", "GITHUB_TOKEN": "should-not-leak"})
monkeypatch.setattr(LocalSandbox, "_get_shell", staticmethod(lambda: r"C:\Windows\System32\cmd.exe"))
monkeypatch.setattr(local_sandbox.subprocess, "run", fake_run)
output = LocalSandbox("t").execute_command("echo hello")
assert output == "ok"
# Platform secrets are scrubbed even on the Windows cmd path (#3861); the env
# is an explicit scrubbed dict, no longer None.
assert calls == [
(
[r"C:\Windows\System32\cmd.exe", "/c", "echo hello"],
{
"shell": False,
"capture_output": True,
"text": True,
"timeout": 600,
"env": {"PATH": r"C:\Windows"},
},
)
]