* feat(extensions): let an out-of-tree extension observe what the agent did
DeerFlow's extension system can contribute middleware, services and routes,
but an extension cannot answer basic questions about a run without reaching
into host internals. Several of the facts it would need are destroyed by the
operations that produce them:
* The middleware chain injects and rewrites a lot of context — date
reminders, recalled memory, compaction summaries, durable-context data,
image payloads, activated skill bodies. Downstream, none of it is
attributable: at the model-call boundary an injected HumanMessage is
indistinguishable from the user's own, and anything wanting to tell them
apart has to pattern-match prompt wording, which breaks on the next copy
edit.
* Two runs of "the same agent" are only comparable if the chain enforced the
same limits, prompts and thresholds. Recovering that from outside means
reading private attributes and guessing which of them change behaviour — a
guess that rots silently as middlewares gain fields.
* The lead-agent factory resolves a model after runtime overrides, renders a
prompt, filters tools through authorization and composes a stack, all
inside one synchronous call, and none of it survives: a middleware sees its
neighbours but not the prompt, the run worker sees a graph but not what
went into it.
* Summarization is destructive by design. N messages leave the context and
one summary enters it; afterwards only the summary exists, so "which
messages became this?" is not reconstructible.
This adds seven neutral facilities so those facts are recorded where they are
still true, and releases the contract package as 0.2.0.
Message provenance
Producers stamp `deerflow_content_kind` / `deerflow_producer_kind` onto the
messages they inject or rewrite. Stamping is unconditional — a fact whose
presence depends on whether an observer is installed is not a fact — and the
keys are server-owned, so provenance cannot be forged from a request.
Middleware self-description
Twelve middlewares declare their own behaviour-affecting parameters through
a duck-typed `release_policy_parameters()`. Long text is hashed rather than
embedded: a declaration is an identity, not a copy of the prompt.
Agent assembly descriptor
`assemble_lead_agent()` returns the graph plus a descriptor whose fingerprint
answers "did anything about this agent change between these two runs?".
`make_lead_agent()` keeps its graph-only signature — it is the LangGraph
Server ABI declared in langgraph.json. Tools and skills are sorted before
hashing because their assembly order is incidental; middlewares are not,
because stack order decides what wraps what. Host build identity is reported
but excluded from the fingerprint, so a redeploy does not invalidate every
agent's identity.
Context compaction observation
Summarization emits the content hashes of the messages it is about to remove
joined to the summary that replaced them. Content is the only identity
available at that seam: the summary does not become a message, and what later
projects it into a request renders it bounded and escaped rather than
verbatim.
Neutral policy, transform and MCP-source facts
Guardrail decisions are published to runtime context under a `__`-prefixed
key; result-rewriting middlewares append a declared, ordered transform trail;
MCP tools carry their credential-free logical origin.
Extension route identity
Contributed routes are session-authenticated and cannot opt out, but
"logged in" and "administrator" are different questions. Extensions get a
neutral projection of the caller rather than the host's auth context, and
`require_admin` fails closed when identity cannot be determined.
Extension-owned tables
An extension that persists data owns its own MetaData and migration chain, so
its tables are absent from Base.metadata and `alembic revision --autogenerate`
proposes dropping them. Extensions declare a table prefix, which is rejected
at registration if it would shadow a host table.
The contract package stays dependency-free and imports no host code; every new
Protocol method has a default so later additions remain additive. The loader's
pre-1.0 rule requires an exact major.minor match, so extensions written against
0.1 are now refused at startup with an actionable install hint rather than
loading into a host that implements a different surface.
uv.lock records the contract package's new version, so `uv sync --locked` still
resolves on a fresh checkout.
* fix(backend): sort gateway service imports
* feat: record guardrail decisions in run events
Persist security-relevant GuardrailMiddleware outcomes (deny and
provider-error fail-open/fail-closed) as middleware:guardrail run
events via RunJournal.record_middleware(), mirroring
SafetyFinishReasonMiddleware. Recording is best-effort: it reads
__run_journal from the runtime context and swallows persistence
failures, so tool execution behavior is unchanged.
The audit payload records tool name/id, agent id, subagent flag, user
role, allow decision, policy id, reason codes/messages, fail_closed
mode, and provider_error flag. Tool input/args and identity fields
(user_id, oauth_*) are deliberately excluded to avoid persisting the
sensitive content being blocked.
The fail-closed provider-error branch returns the denied message
directly from the except block so it records exactly once and does
not fall through to the generic deny branch.
* docs: clarify guardrail journal runtime boundary
* fix: preserve subagent attribution in guardrail events
* test: align guardrail event attribution fixtures
* refactor(guardrails): resolve runtime context once per tool call
Extract `_resolve_context` and thread the already-resolved context dict
through `_build_request` and `_record_guardrail_event` so the
getattr/runtime.context chain runs once per wrap_tool_call instead of twice.
Also document the `is_subagent` field boundary: native subagents do not
inherit __run_journal, so the field is structurally False in persisted
records today; custom runtimes may still supply it with attribution.
Addresses review feedback on #3837 (context-read-twice cleanup and the
is_subagent trade-off note). No behavior change.
Add GuardrailMiddleware that evaluates every tool call before execution.
Three provider options: built-in AllowlistProvider (zero deps), OAP passport
providers (open standard), or custom providers loaded by class path.
- GuardrailProvider protocol with GuardrailRequest/Decision dataclasses
- GuardrailMiddleware (AgentMiddleware, position 5 in chain)
- AllowlistProvider for simple deny/allow by tool name
- GuardrailsConfig (Pydantic singleton, loaded from config.yaml)
- 25 tests covering allow/deny, fail-closed/open, async, GraphBubbleUp
- Comprehensive docs at backend/docs/GUARDRAILS.md
Closes#1213
Co-authored-by: Willem Jiang <willem.jiang@gmail.com>
