Agent Claude Code
Audits error handling for silent failures, broad catch blocks, inadequate logging, unjustified fallbacks, and mock fallbacks in production. Severity: CRITICAL (silent failure / broad catch) / HIGH (poor error message / unjustified fallback) / MEDIUM (missing context). Reports location + severity + hidden errors + user…
Agent Claude Code
Reviews test code for quality, correctness, and adherence to SynthOrg testing conventions: test isolation, mock correctness, markers, parametrize, assertion quality, vendor-name policy, flaky patterns, and web-dashboard test rules. Use for changes under tests/ or web/src//.test..
Agent Claude Code
Verifies Claude Code and OpenCode configuration parity; checks that changes to .claude/ or .opencode/ files maintain dual-tool compatibility.
Agent Claude Code
Analyses type design for invariant strength and encapsulation. Rates Encapsulation / Invariant Expression / Invariant Usefulness / Invariant Enforcement on a 1-10 scale. Flags anti-patterns (anemic models, exposed mutable internals, doc-only invariants, missing constructor validation). Suggests pragmatic improvements…
Hook Claude Code
Runs before the agent uses a tool for Bash, Edit, Write, Edit and Write tool calls, executing check_no_repush_after_failure.sh, check_push_rebased.sh, check_push_throttle.sh, check_ci_before_push.sh, check_no_throttle_override_creation.sh, check_no_baseline_update.sh, check_bash_no_write.sh, check_git_c_cwd.sh…
Hook Claude Code
Runs after a tool call finishes for Edit, Write and Bash tool calls, executing check_web_design_system.py, check_backend_regional_defaults.py, run_edit_time_gates.py, record_push_throttle.sh and rewarm_caches_after_sync.sh via python (5 commands). From Aureliolo/synthorg.
Hook Claude Code
Runs when a session ends, executing run_affected_mypy.py via uv with --stop. From Aureliolo/synthorg.
Settings file Claude Code
Agent settings declaring 3 hook events (PreToolUse, PostToolUse, SessionEnd).
Skill Claude CodeCodex
Analyse application logs from Docker containers: extract, parse, summarize, and detect logging pipeline discrepancies.
Skill Claude CodeCodex
Full PR review pipeline: local agents + external feedback + triage + implement fixes.
Skill Claude CodeCodex
Watch a PR after creation. Polls CI + external reviewer state + open code-scanning/Dependabot/secret-scanning alerts, auto-fixes valid feedback (one push per round), dismisses justified security alerts via API with reason, handles CodeRabbit rate-limit by reposting @coderabbitai review, runs until convergence or…
Skill Claude CodeCodex
Full codebase audit: launches 159 specialized agents (slots 01-159) to find issues across Python/React/Go/docs/website, writes findings to audit/latest/findings/, then triages with user.
Skill Claude CodeCodex
Run one real objective through the whole orchestration loop as an operator, through the dashboard only, recording every collapse before fixing any of it.
Skill Claude CodeCodex
Ship a convention gate: the enforcement script, its wiring, the inventory rows, its tests, and the docs the meta-gate checks.
Skill Claude CodeCodex
Add a settings knob end to end: definition, precedence category, live-apply wiring or a compose-set declaration, and dashboard visibility.
Skill Claude CodeCodex
Plan max-safe parallel worktrees: clean up, survey open issues (excluding renovate/backlog/maybe), map what's inflight, analyse dependencies + file conflicts, and propose the largest conflict-free set of worktrees that each fully close at least one issue. Hands groupings to /worktree setup.
Skill Claude CodeCodex
Run this after squash-merging a PR to clean up the local repo.
Skill Claude CodeCodex
Pre-PR review pipeline: automated checks + review agents + fixes + create PR.
Skill Claude CodeCodex
Research any link, article, tool, or concept and evaluate what it means for this project.
Skill Claude CodeCodex
Review dependency update PRs: changelog analysis, breaking changes, new features, opportunities, and actionable decisions.
Skill Claude CodeCodex
Stand up the local dev arm: the backend built from this worktree running in the operator's own stack, plus a live-reload Vite frontend, with both logs captured.
Skill Claude CodeCodex
Manage parallel worktrees: create with prompts, cleanup after merge, status, tree view, rebase, launch in Windows Terminal tabs.
Command
Analyse Docker container logs.
Command
Full PR review pipeline (local agents + external feedback + triage + fixes).