Embedded runs targeting a CLI runtime provider fall through to the openclaw
harness and call the provider API directly with the runtime's credentials
(cli_runtime_passthrough_openclaw). Anthropic routes direct anthropic-messages
calls on subscription OAuth tokens to metered extra-usage billing — this is
long-standing behavior, not a recent change. Without extra-usage balance,
every such run (e.g. active-memory recall) fails with a billing error; with
extra-usage enabled, the run silently draws paid metered usage instead of the
plan limits the CLI runtime was configured for. Only CLI-backed execution
runs on plan limits for those credentials.
Add an opt-in RunEmbeddedAgentParams.cliBackendDispatch: "subscription-auth"
that dispatches the run through runCliAgent as a one-shot turn when the
provider is claude-cli, a CLI backend is registered, and the ordered auth
profile selection for the passthrough resolves to a subscription (oauth/token)
credential or nothing rather than an API key; resolution stays on stored
credential metadata, with no credential materialization or refresh on the
per-turn path. The dispatch translates toolsAllow into the selectable-backend
surface (native: [], allowlisted loopback MCP tools; wildcard allowlists stay
MCP-only), runs with a fresh CLI process (no live-session reuse; session-
scoped bundle-MCP retirement on run end rather than the process-wide loopback
close), bridges CLI tool result events to onAgentToolResult with native-path
semantics (normalizeToolName + isToolResultError), and drops CLI session
bindings from the result.
The selectable-backend MCP list now also bounds the loopback MCP grant
server-side: the grant carries a per-run gateway tool allowlist enforced in
scoped tool resolution, so tools outside the run's allowlist can be neither
listed nor called even under CLI bypass permission modes where
--allowedTools is advisory.
active-memory recall opts in so recall works on claude-cli subscription-only
instances and stops drawing metered extra usage where it previously could.
Scoped to claude-cli; other CLI runtimes keep the passthrough until their
direct-API contract is verified.
The dispatch also mirrors the run into the run's session transcript through the session
accessor (user turn, tool call/result records as they stream, final assistant
snapshot at run end) so transcript consumers keep parity with embedded runs:
active-memory's persistTranscripts, timeout partial-text salvage, and the
live terminal-search watcher that polls the session file mid-run.
Post-review hardening: canonical anthropic/<model> refs whose configured
agentRuntime is claude-cli resolve through the runtime policy before the
dispatch gate (they previously stayed on the failing passthrough); restricted
dispatches serve an exclusive loopback-only MCP bundle so user/plugin MCP
servers stay outside the run's tool universe; and the transcript recorder
flushes the latest assistant snapshot the moment the run aborts, so timeout
salvage sees partial text even while the killed CLI child is still settling.
Recalls routed to the claude-cli runtime default to a 45s budget (measured
CLI-dispatched runs take 14-20s, over the plain 15s default); explicit
timeoutMs config always wins.
Transcript mirror keeps bare-array tool_result content (claude stream-json
echoes MCP results without a {content} wrapper); dropping it classified every
successful recall as no_relevant_memory.
CLI dispatch resolves inside session/global lane admission so dispatched
runs obey the same lifecycle, placement, and concurrency gates as native
embedded runs.
LOC-ratchet offsets move the loopback grant-context builders to
cli-runner/mcp-grant-context.ts and dedupe the run/prep stage-summary
emitters into attempt-stage-timing.ts; unused type exports dropped and the
now-used stream-message baseline entry removed.
The recall timeout default now consumes the runner's own dispatch
eligibility through a new plugin-runtime seam
(agent.resolveCliBackendDispatchEligibility): API-key and missing-backend
routes keep the passthrough and its plain 15s default.
Eligibility honors an explicitly pinned authProfileId (the credential the
run executes on) before ordered profile selection, in both directions.
Transcript mirror composes buildAssistantMessage + buildUsageWithNoCost
directly; main trimmed the zero-usage wrapper export (ab0ccc244b) before
this branch's usage landed.
Dispatch eligibility is provider-owned: the anthropic plugin's claude-cli
backend declares CliBackendPlugin.subscriptionAuthDispatch and core reads
the registered descriptor instead of a core provider allowlist.
Dispatch fails closed on tool policy (only non-empty named allowlists are
expressible on the CLI surface; deny-all, wildcards, absent allowlists, and
disableTools/modelRun keep the passthrough), threads the pinned
authProfileId into CLI runtime resolution, and emits onExecutionStarted at
the admitted dispatch boundary.
Disable Claude CLI native background Bash and Monitor in OpenClaw-managed print runs, matching the existing fail-closed scheduler-tool behavior. This keeps deferred work on OpenClaw-owned wake paths instead of native Claude callbacks OpenClaw cannot deliver.
Thanks @anagnorisis2peripeteia!
Claude Code built-ins ScheduleWakeup and CronCreate schedule a deferred
re-invocation managed by the persistent CLI runtime. In OpenClaw's
one-shot `claude -p` invocations the process exits at end_turn, so any
wakeup or cron registered during the run has no host to fire into and is
silently lost. Symptom: a CLI session spawns a background sub-agent,
calls ScheduleWakeup to poll for completion, ends the turn, and never
picks up the result — the work finishes unreviewed.
Append `--disallowedTools "ScheduleWakeup,CronCreate"` to both `args`
and `resumeArgs` in the anthropic CLI backend so the model cannot reach
for tools that don't survive the run mode. The right pattern in CLI
sessions is Monitor on the background output file, or a synchronous
sub-agent.
Add `ownsNativeCompaction` capability to CliBackendPlugin so backends
that manage their own transcript compaction (e.g. Claude Code) can
declare it once and OpenClaw defers instead of fighting or failing.
Today only Codex declares compaction ownership (via the embedded runner
path + agentHarnessId). Claude-cli never reaches that path because it
runs as a CLI subprocess with no harness id set, so the safeguard
summarizer fires and hard-fails the turn.
This PR:
- Adds `ownsNativeCompaction?: boolean` to the backend plugin type
- Propagates it through all 4 backend resolution paths
- In `runCliTurnCompactionLifecycle`, when a backend declares ownership
but has no harness endpoint, returns a no-op instead of falling
through to the safeguard
- Sets the flag on claude-cli (first adopter)
Codex's existing native-harness path is unchanged: when
`isNativeHarnessCompactionSession` matches, the harness compaction
endpoint is still called as before.
Generalizes the partial fix in #87785 (codex-scoped) to a capability
any backend can opt into.
Summary:
- Adds a plugin-owned CLI backend argument rewrite hook and wires Anthropic `claude-cli` to translate non-off `/think` levels into Claude Code `--effort`, with docs, changelog, API baseline, and tests.
- Reproducibility: yes. Current main has a high-confidence source reproduction: choose `claude-cli`, set a non ... builds argv from backend args that contain no `--effort` even though `thinkLevel` exists on the run params.
Automerge notes:
- No ClawSweeper repair was needed after automerge opt-in.
Validation:
- ClawSweeper review passed for head be17754009.
- Required merge gates passed before the squash merge.
Prepared head SHA: be17754009
Review: https://github.com/openclaw/openclaw/pull/77410#issuecomment-4372812685
Co-authored-by: stainlu <stainlu@newtype-ai.org>
Derive Claude CLI bypass mode from OpenClaw exec YOLO policy, preserve raw Claude permission-mode overrides, update docs/changelog, and cover global/per-agent policy behavior.
* feat(cli): keep claude cli sessions warm
* test(cli): cover claude live session reuse
* fix(cli): harden claude live session reuse
* fix(cli): redact mcp session key logs
* fix(cli): bound claude live session turns
* fix(cli): reuse claude live sessions on resume
* refactor(cli): canonicalize claude live argv
* fix(cli): preserve claude live resume state
* fix(cli): close dead claude live sessions
* fix(cli): serialize claude live session creates
* fix(cli): count pending claude live sessions
* fix(cli): tighten claude live resume abort
* fix(cli): reject closed claude live sessions
* fix(cli): refresh claude live fingerprints
* fix(cli): stabilize MCP resume hash
* fix: preserve claude live inline resume (#69679)
---------
Co-authored-by: Frank Yang <frank.ekn@gmail.com>
Stop injecting CLAUDE_CODE_PROVIDER_MANAGED_BY_HOST into Claude CLI runs and strip inherited/backend overrides before spawn.\n\nAlso repairs the Zalo setup allowlist prompt wiring needed by the current main check gate.\n\nThanks @Alex-Alaniz.
The Claude CLI backend uses `--output-format json`, which produces no
stdout until the entire request completes. When session context is large
(100K+ tokens) or API response is slow, the no-output watchdog timer
(max 180s for resume sessions) kills the process before it finishes,
resulting in "CLI produced no output for 180s and was terminated" errors.
Switch to `--output-format stream-json --verbose` so Claude CLI emits
NDJSON events throughout processing (init, assistant, rate_limit, result).
Each event resets the watchdog timer, which is the intended behavior —
the watchdog detects truly stuck processes, not slow-but-progressing ones.
Changes:
- cli-backends.ts: `json` → `stream-json --verbose`, `output: "jsonl"`
- helpers.ts: teach parseCliJsonl to extract text from Claude's
`{"type":"result","result":"..."}` NDJSON line
Note: `--verbose` is required for stream-json in `-p` (print) mode.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Vito Cappello