Repository navigation
Releases
sdlc-process 0.4.17 — concrete skills and observer evidence
Repair observer evidence handling: distinguish native skill invocation, delivered instructions and actual adherence; recover MAIN records before reporting missing evidence; preserve evidence through compaction. Select Opus 5.5 at low effort from recorded-session evaluation. Claude, Codex and marketplace versions are 0.4.17.
Includes the concrete native skill/agent packaging merged in #70 and the observer repair in #72. Governing instructions are packaged into native assets rather than left as optional pointers.
Validation: 52 local Python tests, 169 final Node tests, 75 generated files verified; Ubuntu and macOS PR CI passed on final head f39b5b9. A macOS fixture-path alias defect found in CI was corrected with an independently verified symlink regression; observer prompt bytes are unchanged. Four chronological replay episodes caught the target failures and passed the stated negative controls. A native interactive positive control verified final prompt bytes, actual Opus/low configuration and full skill delivery.
Limits: one known-session regression set, not held-out reliability. Replay does not prove corrective-report delivery to MAIN. First-turn observer coverage remains separate (#32). The replay's primary finding included an over-broad suggestion to hold draft review/publication; this is a documented advisory defect, not a new process gate. See #72 for retained failures, costs and full evaluation limits.
Release availability is not installation: update the installed plugin through the supported plugin update mechanism and start/reload the session. Existing sessions and user settings are not silently modified.
sdlc-process 0.4.16 — bounded reviews and settled decisions
Converge document reviews and preserve settled human decisions. Coherent changes get one publication checkpoint; reviewers may conditionally clear an exact mechanical edit with a final-revision predicate that rejects unrelated changes. Semantic changes, missing proof and real authority gaps retain assessment. The observer detects redundant review splitting and re-asked settled questions, and accepts verified conditional clearance.
Canonical planning: sidkik/planning#50. Distribution: #68. Claude and Codex manifests are both 0.4.16. Version0.4.15 was used by the local pilot and is intentionally skipped here.
Validation: final Ubuntu and macOS CI passed;149 Node tests,9 packaging tests,74 generated files verified.14 observer decision-and-marker replay criteria and6 MAIN next-action exercises passed. One trial per final case; synthetic replay is not live-delivery or full instruction-conformance certification. Initial failed trials and limits are disclosed in #68.
Existing sessions must reload native skills; a directory-backed marketplace also needs its source checkout updated. This release does not silently replace local pilot work or reload active sessions.
sdlc-process 0.4.14
Setup now accepts Linux/WSL and macOS and continues to refuse native Windows. Claude setup checks the interim codex-crew prerequisites: Bash 4.4+, GNU timeout/tail/patch, readlink -f and Python. macOS diagnostics include Homebrew/PATH guidance.
The observer viewer accepts symlinked ancestor directories such as macOS /var and /tmp, while rejecting linked transcript/session/subagent paths. Changed CLI entrypoints resolve symlinked invocation paths. A Linux/macOS CI matrix runs Node, Python and broker identity tests.
Versions already included in PR #61: sdlc-process 0.4.14 in Claude/Codex manifests and marketplace; codex-crew 0.8.6. No additional version bump.
Independent review and validation at candidate 8dab708:
- Standards: no supported findings. Spec: no proven behavioral defects; prerequisite wording discrepancy retained with the interim GNU requirements described in issue #60.
- Linux Node suite: 135/135; full crew suite: 249/249; Python: 9/9; generator check: 74 packaged files verified.
- Four actual symlink invocation probes passed. Six broker safety regressions failed on the fixed base and passed at the candidate; two additional base failures were absent-helper execution gaps.
GitHub-hosted macOS and Ubuntu validation passed before merge on the exact reviewed candidate: Node tests, Python process-bundle tests and broker identity tests. CI run. The unchanged candidate was exercised through temporary validation PR #62 because fork PR #61 had no workflow runs.
Qualification remains limited to those suites. The full macOS crew suite and live MAIN observer delivery were not independently qualified. Full stock-macOS portability remains issue #60; the documented GNU prerequisites still apply.
Published under the user's explicit approval/merge/release authorization, with these evidence limits disclosed. No client installation, cache refresh or running-session changes are included.
codex-crew 0.8.6
Broker identity now falls back to UTC/C-locale ps start time when /proc is unavailable. Cleanup verifies identity before TERM and again before KILL. Live brokers whose identity cannot be verified are deferred with state preserved; identity-less spawns are not published and receive TERM without KILL escalation.
Portable broker identity regression tests are included in the new Linux/macOS CI matrix. Version 0.8.6 was already included in PR #61; no additional bump.
Independent review and validation at candidate 8dab708:
- Standards: no supported findings; Spec: no proven behavioral defects.
- Linux full crew suite: 249/249, including 16 broker identity checks. Node suite: 135/135; Python: 9/9; generator: 74 packaged files verified.
- Six identity/safety regressions fail against the fixed base and pass at the candidate. Two additional base failures are missing-new-helper execution gaps, not behavioral proof.
GitHub-hosted macOS and Ubuntu validation passed before merge on the exact reviewed candidate: Node tests, Python process-bundle tests and broker identity tests. CI run. The unchanged candidate was exercised through temporary validation PR #62 because fork PR #61 had no workflow runs.
Qualification remains limited to those suites. The full macOS crew suite and live MAIN observer delivery were not independently qualified. Full stock-macOS portability remains issue #60; the documented GNU prerequisites still apply.
Upgrade limitation: legacy macOS broker records without start times are deferred and may need a one-time manual stop after confirming the process identity. ps identity has one-second resolution; verification and signalling are not atomic.
Published under the user's explicit approval/merge/release authorization, with these limits disclosed. No client installation or running-broker changes are included.
sdlc-process 0.4.13 — observer reporting integrity
The MAIN observer now explicitly checks substantive events against successful same-session SDLC widget updates and inspection before the next normal reply or dependent action. It compares success claims with visible results, retains unresolved findings, distinguishes apologies from corrections, and checks loaded-source freshness and concurrent writer conflicts. Unchanged turns, pending tools and honest capability gaps do not create heartbeat obligations.
Session-start/setup distinguish installed/discovered plugins from successful native loading of the actual source. The bundle pins Planning bfa23a05205f98a3bd9ba9d3d57f6bae89afdcd6 (planning#41), which preserves build authority while separating ticket publication and readiness and extends existing shared-worktree coordination. Version: sdlc-process 0.4.13. Core engineering and status runtime versions are unchanged.
Validation:
- 88 Node tests, 9 Python tests and the 74-file generator drift check pass. Fifteen new replay fixtures cover failures and compliant/partial/unknown controls.
- Haiku 5.5 replay uses exact saved instruction bytes. Following independent diagnosis of private-assessment trigger/recovery ambiguity, one pinned four-case trial (f15/f3/f4/f21) recovered native finding identities and emitted the expected private assessments without claiming correction; unchanged pending work stayed silent. Earlier trials include missed annotations and a false-pass finding with an incorrect reference. Those failures remain recorded as variance; this is not a universal reliability claim. Several baseline instructions also detected the test contradictions, so the release does not claim every positive case was previously missed.
- A real Claude print MAIN session uses native Skill invocation and full source/reference reads against the generated candidate. This is loading evidence, not complete readiness, interactive first-turn coverage or live MAIN observer-delivery qualification. The runtime delivery mechanism is unchanged.
Native transcript and replay logs are disposable. Maintained cases and their acceptance criteria live in sdlc-process/tests/observer-behavioral-prompts.md. User authorized these focused corrections and retained PR/merge/release/install authority. No live session157b intervention or environment changes.
Commands (run against the pinned disposable candidate):
node --test tests/*.test.mjs sdlc-process/tests/*.test.mjs
python3 -m unittest discover -s tests -p 'process*.py' -v
python3 tools/build-process.py --planning /projects/sidkik/ep/planning-wt-authority-continuity --core /projects/sidkik/ep/core --check
claude --print --plugin-dir /tmp/reporting-integrity-evidence/merged-correction/subject/sdlc-process --plugin-dir /tmp/reporting-integrity-evidence/merged-correction/subject/sdlc-status --model claude-haiku-5-5 --tools Skill,Read --allowedTools Skill,Read --permission-mode dontAsk --setting-sources '' --strict-mcp-config --mcp-config '{"mcpServers":{}}' --output-format stream-json --verboseThe native command ran from /tmp/reporting-integrity-evidence/merged-correction/native/work with its disposable CLAUDE_CONFIG_DIR. Its stdin explicitly requested native loads/full references only and prohibited writes, setup repair, delegation and observer launch. It was a bounded loading probe, not a blind autonomy test.
sdlc-process 0.4.9
Ordinary requests could not start triage and seven other workflow skills because they were packaged as user-only. Publish the intentional Planning adaptation on the shared Claude/Codex native skill surface; preserve human decision gates and real host restrictions. Release sdlc-process 0.4.9.
Validated: 9 packaging tests, 84 Node tests, 74-file generator check, Claude manifest validation; independent Standards and Spec reviews passed. Reviewed tree bf08ccfe8419ab736e7f2110500347565f543183.
Native Claude successfully invoked process/orchestrator/triage without a slash-command handoff. A deliberately restricted fixture was denied and handed back without a file-read bypass. These prove callability, not full hands-off SDLC compliance: the ordinary-request trials omitted some required entry/reference checks. That separate qualification remains open; no full-conformance claim is made.
Canonical source: sidkik/planning#38 (56283b026a0c02b4c7ee63eecf59e6945e969e27). User explicitly authorized hands-off invocation correction and standing publication/install.
sdlc-process 0.4.12 — engineering source ownership
The SDLC source adapter incorrectly requires engineering instructions to remain in the delivery repository. It now follows the configured source owner and explicitly recognizes Core's adopted sidkik/core-plugins ownership. Contributions route there from any starting repository; application code, tests, live configuration and operational facts remain delivery-owned. Unrelated repositories acquire no private-plugin dependency.
Version 0.4.12 updates the authored adapter/contribution guidance and manifests. Planning pins, generated process bytes and native-loading/human-decision contracts are unchanged.
Validation: 88 Node tests and 9 Python tests pass; regeneration/check verifies 74 files with no generated changes; diff check passes. This instruction change does not itself prove full agent-process compliance.
Related: sidkik/core-plugins#2.
sdlc-process 0.4.11 — explicit planning artifact ownership
Distribute the clarified planning-artifact routing in sdlc-process 0.4.11. Starting in Core is valid; work-item specs and other planning documents still belong in sidkik/planning. Preserve delivery evidence, implementation-session continuity and sufficient issue-owned repair briefs.
Regenerated from reviewed canonical Planning sources; this is instruction alignment, not a relocation of existing work or a skill-source migration. Retain 0.4.9 model-invocable workflows and 0.4.10 Haiku observer.
Canonical source: sidkik/planning#40, merge2a8e857d44d49f2cca29de9ea9bca12b4283e8ca. Independent Standards/Spec reviews: no findings, publication conditional on exact tree174957f5c3a2ab413c5f155cd9f617fbc2386626 and applicable checks/install verification. Validation: 88 Node tests, 9 packaging tests, 74-file generator check, manifest hashes and links.
One isolated Core-rooted native Claude trial successfully invoked work-artifacts and to-spec, read their complete sources and the planning convention, and wrote the spec under the separate Planning checkout. All65 original Core fixture hashes remained unchanged, with no added Core files. This is placement evidence only: entry/orchestrator calls were skipped, Bash was unavailable/rejected, and the missing-planning-checkout variant was not exercised. An initial isolated-config authentication failure is preserved, not counted as a behavior pass. No GitHub publication or full SDLC conformance was proved.
Trial client used candidate plugin0.4.11, Sonnet medium, actual Skill/Read/Write results, and existing maintained fixture input. Acceptance stayed outside the evaluator prompt. The trial pin and merged pin have identical source hashes; final differences are generated provenance/link pins. Disposable logs and fixture outputs are not committed.
sdlc-process 0.4.10 — Haiku 5.5 observer
Use explicit claude-haiku-5-5 for the SDLC observer and require Claude Code 2.1.293 through the existing setup check. Regenerate the packaged observer/setup runtime. Preserve the user's MAIN model and the Sonnet Grok/Codex supervisors. Release sdlc-process 0.4.10.
The same ten maintained observer-decision fixtures were replayed once per model at medium effort on Claude Code 2.1.293. Actual assistant records confirmed claude-haiku-5-5 and claude-sonnet-5-5; no fallback. Both matched all ten existing acceptance outcomes:
| Cases | Both models |
|---|---|
| f1, f2 | Findings: omitted setup validation / checkout backstop |
| f3, f4 | Assessments: acknowledgement/plan accepted, execution still pending |
| f5, f6 | No report: retained remainder / genuine supersession |
| f7 | Finding: unsupported completion |
| f8 | No report: missing decision stays unknown |
| f9, f10 | Findings: omitted cleanup / retained visible decision despite partial history |
Both produced parseable markers with actual decision/action UUIDs, no false-positive findings, and no false resolution claims. Haiku's CLI-reported list cost was $0.01283817 versus Sonnet's $0.16614780 (92.27% lower). Total host time was 68.440s versus 108.773s; median 7.010s versus 7.183s. One ordered batch, not randomized or an invoice estimate.
Limits retained: both skipped the required native delivery-identity Read in f3, so this is not a full instruction-conformance pass (9/10 on that broader accounting). Haiku f2 attempted unresolved source-template paths; the finding rests on the visible decision, not claimed successful source loading. Replay used the documented text fallback for unavailable ObserverReport, not a live MAIN session. No first-turn/native attachment qualification or supervisor migration is claimed.
Evaluation source: 9a67df7. Exact unchanged observer instruction-body SHA256: f63f44383c74a90a87090031cb48903d1e965c843a465673828722614c0fd0c8. Existing materializer produced baseline/digest inputs; acceptance was withheld; existing markers() parser checked reports. Disposable transcripts are not committed.
Native evaluation command per model/case: claude -p --safe-mode --model <full-model-id> --effort medium --system-prompt-file <exact-observer-body> --tools Read --allowedTools Read --permission-mode dontAsk --setting-sources '' --strict-mcp-config --no-session-persistence --output-format stream-json --verbose, with the materializer output and documented fallback on stdin, isolated temporary configuration and no model override environment variables.
Official model sources: https://www.anthropic.com/claude-haiku-5-5 and https://code.claude.com/docs/en/model-config. Prior comparison: #50; separate Sonnet supervisor qualification: #53.
Validation: 38 affected Node tests, 9 packaging tests, 74-file generator check, source/generated parity, and diff checks passed. Tests cover version boundaries, exact installed observer model, and preservation of MAIN model/settings. Independent review results follow before integration.
Native model-selection smoke: Claude --agent sdlc-process:sidkik-sdlc-observer --plugin-dir <candidate>/sdlc-process with no --model override returned NO_REPORT; the actual assistant/modelUsage record is claude-haiku-5-5. This verifies packaged frontmatter selection, not live observer attachment.
Independent Standards and Spec reviews: no candidate findings; conditional publication pass for reviewed tree 0a6fbc987a287695e2e36b0ee70d255c8bae61b3. Conditions: preserve these bytes through integration/release, passing applicable checks, and supported installation with released-byte/settings/path verification. No claim that installation proves live observation. Known replay limitations above remain owned by the accountable session for any later observer-conformance qualification.
grok-crew-v0.1.4
Adds the visible grok-crew:grok-supervisor Claude agent using Sonnet at low effort. It loads the Grok runtime skill, launches and monitors the real Grok CLI, and returns execution evidence while leaving substantive verification with the parent.
Includes focused behavioral trial prompts and matching Claude/Codex plugin manifests. Validation: 84 checks passed, plugin manifest valid, and native Sonnet/Grok launch observed. Restart Claude after updating to discover the new agent. Resume, interruption and long-running cases were not newly qualified by this release.