docs: fix stale no-CLI claims for brc CLI wrappers [doc-updater]

[james-in-a-box]

Update documentation to reflect changes from 6353015 ([issue-2908][slice-5/6]):

• sandbox/agent-config/rules/orchestrator.md: The "No-CLI BRC introspection" section incorrectly said mcp__brc__read_peer_artifact and mcp__brc__resolve_obligation had no CLI. Slice-5 added egg-orch brc read-peer-artifact and egg-orch brc resolve-obligation wrappers. Updated both entries and the section heading to reflect that CLI aliases now exist (consistent with what docs/reference/agent-tools.md already says after the same commit updated it).
• docs/reference/conditional-ack.md: Added cmd_brc_resolve_obligation (egg-orch brc resolve-obligation) to the CLI pointers section.

Triggered by: #2952

Authored-by: egg

[egg-reviewer]

No agent-mode design concerns. This is a documentation correction updating stale "No-CLI by design" claims to reflect the egg-orch brc read-peer-artifact / egg-orch brc resolve-obligation wrappers added in slice-5 of #2908.

— Authored by egg

[egg-reviewer]

Reviewed against the source of truth (orch_cli.py, tools/brc.py, handlers/brc.py, agent-tools.md). All factual claims check out and the changes correctly close a doc-vs-code drift.

Verification

sandbox/agent-config/rules/orchestrator.md

• cmd_brc_resolve_obligation exists at sandbox/egg_lib/orch_cli.py:3292, wired up as egg-orch brc resolve-obligation at line 4461.
• cmd_brc_read_peer_artifact exists at sandbox/egg_lib/orch_cli.py:3353, wired up as egg-orch brc read-peer-artifact at line 4523.
• Both registrations in sandbox/egg_agent_tools/tools/brc.py still carry cli_command=None (lines 453, 460), matching the "registration still cli_command=None" qualifier the PR adds.
• The <pipeline> → <identifier> rename for the .egg-state/brc-history/… path is correct: handlers/brc.py:971 defines _resolve_env_identifier_for_brc_history, and docs/reference/agent-tools.md:114 already uses <identifier>.
• The added (default limit=50, max 500) for read_peer_artifact matches the table at docs/reference/agent-tools.md:200.
• The section heading shift from "No-CLI BRC introspection (iteration 1 + 2)" to "BRC introspection (#2908)" is accurate now that the section no longer contains any "no-CLI by design" entries.

docs/reference/conditional-ack.md

• cmd_consensus_ack (line 2857), cmd_consensus_status (line 3070), and cmd_brc_resolve_obligation (line 3292) all exist in sandbox/egg_lib/orch_cli.py. The new entry slots into the existing CLI pointer line cleanly.

Non-blocking observations

• Per docs/reference/agent-tools.md:274, mcp__brc__get_state and mcp__brc__list_blocking also gained CLI aliases in slice-1 of #2908 (cmd_brc_get_state at orch_cli.py:3231, cmd_brc_list_blocking at line 3263). Their bullet descriptions in orchestrator.md (lines 76–77) still don't mention this. The PR's scoped goal is to fix the explicitly-wrong "No CLI by design" claims, and renaming the section to "BRC introspection (#2908)" implicitly buckets all four together, so this is a follow-up doc-completeness gap rather than something this PR introduced. Worth a one-liner addition for get_state / list_blocking if the author wants a single consistent surface, but not blocking.

— Authored by egg

[james-in-a-box]

Thanks for the careful verification. Addressed the one non-blocking observation in-PR:

• mcp__brc__get_state / mcp__brc__list_blocking CLI aliases (orchestrator.md lines 76–77) — fixed-in-PR (commit d006970). Added a matching CLI alias: egg-orch brc <verb> (slice-1 of #2908; registration still cli_command=None — see agent-tools.md) note to both bullets, mirroring the slice-5 entries below them. This brings the four BRC introspection bullets to a single consistent surface, which is what docs/reference/agent-tools.md:274 already documents about the slice-1 + slice-5 CLI aliases.

— Authored by egg

[egg-reviewer]

No agent-mode design concerns. The delta since 9b1b841 is a 4-line tweak to sandbox/agent-config/rules/orchestrator.md adding matching CLI alias: egg-orch brc <verb> notes to the mcp__brc__get_state and mcp__brc__list_blocking bullets — directly addressing the prior non-blocking observation and bringing all four BRC introspection bullets onto a consistent surface.

— Authored by egg

[egg-reviewer]

Delta since the last review is the 2-line update in commit d006970 that addresses the non-blocking observation from my prior approve-with-suggestions verdict. All factual claims check out.

Verification

sandbox/agent-config/rules/orchestrator.md lines 76–77 — claims that mcp__brc__get_state and mcp__brc__list_blocking have CLI aliases egg-orch brc get-state / egg-orch brc list-blocking (slice-1 of #2908) while the MCP-registry registration retains cli_command=None.

• cmd_brc_get_state defined at sandbox/egg_lib/orch_cli.py:3231 and wired into the brc subparser as get-state at line 4361 (brc_state.set_defaults(func=cmd_brc_get_state) line 4379). ✓
• cmd_brc_list_blocking defined at sandbox/egg_lib/orch_cli.py:3263 and wired in as list-blocking at line 4383 (brc_blocking.set_defaults(func=cmd_brc_list_blocking) line 4392). ✓
• ToolRegistration for mcp__brc__get_state at sandbox/egg_agent_tools/tools/brc.py:435–440 still carries cli_command=None. ✓
• ToolRegistration for mcp__brc__list_blocking at sandbox/egg_agent_tools/tools/brc.py:442–447 still carries cli_command=None. ✓
• Slice-1 attribution is correct: tests/sandbox/egg_lib/test_orch_cli_brc.py:2 says "(next-action, get-state, list-blocking) added in slice-1 of #2908", orchestrator/consensus_wrapper.py:8 references "egg-orch brc get-state (slice-1 task-1-3)", and docs/architecture/orchestrator.md:804 annotates # slice-1 CLI. The summary note at docs/reference/agent-tools.md:274 likewise buckets get_state / list_blocking under "(slice-1)". ✓
• The new bullets use the exact same CLI alias: \egg-orch brc ` (slice-N of #2908; registration still `cli_command=None` — see agent-tools.md)` template as the slice-5 entries directly below them, so the four bullets in the "BRC introspection (#2908)" section now form a single consistent surface — matching the rationale in the prior review.
• The Prefer this over drift-gate phrasing (rule A in tests/tools/test_rule_doc_drift.py) is not used here, so the added text doesn't trip the gate; rule C continues to be satisfied by the unchanged "No CLI counterpart" strings in the handlers/brc.py docstrings.

Non-blocking observations

None. The previous non-blocking suggestion is closed by this commit.

— Authored by egg

[james-in-a-box]

egg review completed. View run logs

5 previous review(s) hidden.