Commit Graph
9 Commits
Author SHA1 Message Date
Mike Schwarzanddev-qa 57380a250e Release0.6.0: works for strangers (#109)
* chore(release): prepare OpenRig 0.6.0

Add per-seat permission selection and typing protection, rig-wide Herdr entry,
Node 22/24 support with SQLite 13, and first-use improvements. Label the optional
Jev/Rig Stream classifier and Slack manifest setup assistance experimental.

* docs: finalize 0.6.0 verification scope and release references

Record the accepted Node 22 candidate install, retained Node 24 evidence,
and unperformed native permission checks without broadening test claims.

---------

Co-authored-by: dev-qa <dev-qa@openrig.local>
2026-09-28 14:07:36 -07:00
Mike SchwarzandOpenRig contributors 5c54703035 Release 0.5.16: choose CLAUDE.local.md for Claude managed blocks (#25) (#63)
* feat(daemon): let a rig choose CLAUDE.local.md for Claude managed blocks (#25)

A rig spec can now declare

    managed_blocks:
      claude-code: CLAUDE.local.md

so OpenRig writes its managed instruction blocks for Claude Code members to
CLAUDE.local.md instead of a git-tracked CLAUDE.md. The default stays
CLAUDE.md and Codex members stay on AGENTS.md.

- Validation accepts only the claude-code key with CLAUDE.md or
  CLAUDE.local.md, and rejects anything else before a member launches.
- The selection is stored on the rig (migration 085) and bound once in
  StartupOrchestrator.startNode, so launch, restore replay, relaunch,
  continue and added members all write the same file. Startup
  guidance_merge, profile managed_block projection and the conflict
  target all use it.
- Teardown cleans only the selected file. The other file is never edited,
  moved or deleted; blocks already written to CLAUDE.md stay until the
  user removes them, as the reference docs explain.
- Export and the bundle rig.yaml rewrite round-trip the field.

Reported and proposed by @hvpaiva.

* docs(daemon): say accurately which paths deliver the Claude managed-block file

Handover launches the successor directly and writes no guidance, so the
startNode and migration comments no longer list it among the delivery paths.
Comment-only change.

* release: prepare 0.5.16

Set the version to 0.5.16 and finalize the changelog and release notes for
the Claude instruction-file option (#25), the advisory portability report (#56)
and the README update (#54).

* docs: keep uncommitted edits safe when removing old Claude blocks

Recommend deleting only the managed sections by hand, and warn that
git restore CLAUDE.md discards every unstaged change to the file. Say that
portability findings never fail the check while operational errors still do.

---------

Co-authored-by: OpenRig contributors <noreply@openrig.dev>
2026-09-26 02:05:03 -07:00
v-openrig-build 7eef424446 docs: explain first-user permission choices and launcher precedence 2026-09-24 17:53:40 +00:00
v-openrig-build c8afda612c docs: repair public reference provenance and discovery 2026-09-09 11:10:02 +00:00
v-openrig-build c8f0e98f81 fix(startup): retire lean proof before readiness checks 2026-09-07 04:30:05 +00:00
v-openrig-build 213e5a56e5 Honor authored startup proof selection 2026-09-07 04:30:05 +00:00
v-openrig-buildandClaude Fable 5 fbacbe8010 docs(policy): ship the custom-example policy + rig-spec touchpoint (4.8 carry)
Lane-A carry (qitem 4b805845; PM scope ruling option A — the user-authorable
path gets a worked example; authored by review-r1, delivered via acting-orch,
byte-verified 1ebf873b / resulting docs fb03335f).

- packages/daemon/policies/examples/my-cautious-dev.policy.md: placement
  policies/examples/ is the censused minimal-sibling convention (no prior
  examples convention existed; PM has path-flexibility if they rule otherwise
  at acceptance). Ships via build-package's recursive policies staging
  (cp -r), and is structurally NOT materialized as a builtin — the
  materializer's known-four allowlist admits only locked|standard|open|yolo
  (builtin suite re-run green beside this carry).
- docs/reference/rig-spec.md: the delivered touchpoint patch applied;
  resulting file hash matches the delivery manifest.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-05 03:04:00 +00:00
Mike SchwarzandClaude Opus 4.6 18d66e3a70 Add optional docs field to RigSpec (#8)
* Add optional docs field to RigSpec for bundled documentation

Add a top-level `docs` array field to the pod-aware RigSpec schema.
Each entry has a `path` field (safe relative path, same validation as
culture_file and startup file paths).

The engine does not consume these files for instantiation — they are
metadata for humans and agents setting up the environment before launch.
The bundle packer includes listed docs in the archive so setup
instructions travel with the distributable artifact.

Schema changes:
- types.ts: add RigSpecDoc interface and docs field to RigSpec
- rigspec-schema.ts: validate docs array with path safety checks
- rigspec-schema.ts: normalize docs into typed array
- rigspec-codec.ts: serialize docs field

Bundle changes:
- pod-bundle-assembler.ts: collect docs files into the bundle archive

Tests:
- rigspec-schema-docs.test.ts: 7 tests covering validation, rejection,
  normalization, and path safety

Reference doc:
- docs/reference/rig-spec.md: document the docs field

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* Fix review findings: null guard, missing-docs failure, bundle tests

Addresses all findings from R2 and orch1-peer independent reviews:

1. Guard against null/non-object entries in docs validation — docs: [null]
   and docs: [42] now return clean validation errors instead of crashing
   the validator (R2 #1, peer #1)

2. Bundle assembly fails loudly when a declared docs file is missing from
   disk — prevents silently shipping bundles that advertise docs they
   don't contain (R2 #2, peer #2)

3. Add bundle-level tests proving docs are collected and missing docs
   fail assembly (R2 #3)

Tests: 5 new tests (3 schema null/primitive/string + 2 bundle positive/negative)
Full suite: 128 files, 1722/1722 passing

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* Provision runtime config and disclose setup mutations

* Fix: force acceptEdits for managed sessions, don't inherit restrictive modes

deliverStartup unconditionally sets permissions.defaultMode = "acceptEdits"
for managed Claude sessions instead of preserving existing restrictive modes
via nullish coalesce. A pre-existing denyAll or other restrictive mode in
.claude/settings.local.json would previously survive provisioning, causing
managed rigs to hit prompt/block behavior in reused workspaces.

Found by: dev1-qa independent repro
Fix: change ?? "acceptEdits" to unconditional "acceptEdits" assignment
Test: regression test seeds denyAll, verifies deliverStartup overrides it

Full suite: 128 files, 1723/1723 passing

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* Fix JSON normalization, disclosure copy, and README accuracy

Addresses CodeRabbit findings verified by QA:

1. readJsonObject now excludes arrays (typeof "object" includes arrays
   in JS). Prevents silent property loss when settings.local.json or
   .mcp.json contains a bare JSON array instead of an object.

2. provisionPermissionsAndMcps refactored to use readJsonObject,
   readJsonObjectField, and readStringArray consistently instead of
   raw JSON.parse + manual type checks. Same helpers used elsewhere
   in the adapter.

3. Setup disclosure for .claude/settings.local.json now mentions
   statusLine mutation alongside permissions (CodeRabbit item 3).

4. README clarifies that managed runtime boot performs config writes
   while rig setup only discloses the paths (CodeRabbit item 6).

5. Regression test: array-valued settings.local.json and .mcp.json
   are recovered to proper objects during provisioning.

Full suite: 128 files, 1724/1724 passing

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

* Harden setup disclosure tests against copy churn

---------

Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-11 18:42:35 -07:00
Mike SchwarzandClaude Opus 4.6 cfcd86ee1c Move reference docs to docs/reference/ as canonical source of truth
- Move reference docs from packages/daemon/docs/reference/ to
  docs/reference/ at the repo root — this is the authoring location
- Update .gitignore to track docs/reference/ alongside docs/as-built/
- Update build-package.sh to copy from docs/reference/ into the
  assembled CLI package (not from inside the daemon package)
- Daemon startup still copies the bundled docs to ~/.openrig/reference/
  on start for stable agent-accessible paths

Flow: docs/reference/ (edit here) → build → cli/daemon/docs/reference/
(npm package) → daemon start → ~/.openrig/reference/ (agents read here)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-11 13:55:43 -07:00