v1.4.0 - #5
Merged
Merged
Conversation
Normalize *.sh/*.json/*.md to LF and run-hook.cmd to CRLF so hooks work identically on Windows (Git Bash) and Unix. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Move blocking gates (architecture-ready, token-budget) to PreToolUse(Task) with JSON permissionDecision+exit 0; keep the commit gate on PreToolUse(Bash) with an in-handler 'if' and a stdin guard. Bug pipelines gate on debug-report.md instead of architecture.md. Track active agents as a set (add clear-active-agent.sh on SubagentStop) and anchor SubagentStart to this plugin's own agents. Normalize Windows paths in the ownership check; unify the deny convention; drop fixed /tmp paths for mktemp. Anchor APPROVED to the Approval-status section. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
archive-task.sh moves the previous task's file bus to history/<ts>/ (keeping .gitignore and the .codex-availability cache) so stale artifacts can't leak into the next task's gates. init-claude-team.sh no longer early-exits and reports created/repaired/ok per resource. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Materialize codex-code-reviewer and codex-doc-reviewer (color: yellow) from the build spec, with fail-closed codex exec wrapping and fallback to the internal Reviewer. Mark the build-codex-reviewers spec as done. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Namespace bundled MCP tools (mcp__plugin_dev-team-agents_*); add analyst [VALIDATE-PLAN] mode and reviewer PLAN-REVIEW mode; give git the Edit tool and a fast/trivial/setup commit carve-out; orchestrator spawn-fallback rule and archive step; devops owns memory/project.md; narrow brainstorm/clarify triggers; revert analyst and architect to model: opus (lift the temporary fable pin). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Role-scope the brainstorming/clarifying skills to the Analyst; fix offset/limit usage, the Bash(*) allowed-tools token, and broken cross-links; refresh the operating manual roster, ownership table, and HITL lists. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Those belong in the user config, not the plugin manifest. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Correct the enforcement-hook count to nine (add clear-active-agent), refresh the roster/commands/roadmap, and drop the temporary fable note. Set the canonical GitHub org to app-builders-club and unify author/attribution to App Builders Club across plugin.json, marketplace.json, LICENSE, and README. Record the audit and fix plan in AUDIT-FIX-PLAN.md. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
scripts/smoke-test.sh checks shell syntax, hooks.json structure and event keys, ownership validation on Windows-style paths, agent-reference integrity, agent frontmatter, and version consistency across plugin.json/CHANGELOG/README. .github/workflows/ci.yml runs it on push/PR plus an advisory shellcheck. Extend .gitattributes to keep *.yml/*.yaml at LF. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
orchestrator: token budget is per-session and advisory (warn 80% / deny new dispatch 100%), not per-task. ownership validator: remove the unused .runtime/ allow-rule. codex-detect: clarify that .codex-availability is a decision record rewritten on each probe (not a skip-probing cache). run-hook.cmd already prefers Git bash before PATH lookup; sessions 'unknown' fallback accepted as benign. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Three orchestrator-facing methodology skills formalizing bus mechanics: the task lifecycle and 4-status routing, safe parallel-developer merge (worktree + per-task-file isolation), and the spawn-failure fallback ladder. Wired into orchestrator frontmatter; skill counts bumped to 19 (18 methodology + manual) in README/INSTALL/plugin.json. Smoke test now checks every skills/*/ has a SKILL.md and every agent skills: reference resolves. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Introduce hooks/validators/ownership.json as the declarative single source of truth for .claude-team ownership (21 agents, allow/deny fixtures). The pure-bash validator stays the runtime enforcer (no JSON parsing on the security path); scripts/smoke-test.sh now verifies the validator's agent set matches the spec, drives 42 behavioral allow/deny fixtures through the validator, and checks every owning agent appears in the operating-manual table. Deliberately favors verification over runtime codegen — see AUDIT-FIX-PLAN.md E1. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Post-audit hardening release. Cut the [Unreleased] section to [1.4.0] (2026-07-12) with updated compare links, bump plugin.json, and update the README status line and roadmap. New this release: CI smoke test + workflow, ownership single-source-of-truth spec, and the task-lifecycle/parallel-merge/spawn-fallback skills, on top of the audit fix batch. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.