메뉴
Use when approving, rejecting, or marking implemented Herald GEPA pending approval items, including requests like "approve GEPA changes", "approve all pending GEPA", or approval of a specific queue key from docs/superpowers/gepa-pending-approvals.md.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
| name | approve-gepa-changes |
| description | Use when approving, rejecting, or marking implemented Herald GEPA pending approval items, including requests like "approve GEPA changes", "approve all pending GEPA", or approval of a specific queue key from docs/superpowers/gepa-pending-approvals.md. |
Use this skill for Herald's GEPA approval lane. Approval records a user's decision about workflow suggestions recovered from autopilot self-reflection; it does not implement the approved improvement unless the user separately asks for implementation.
Use the existing Herald autopilot queue tooling:
docs/superpowers/gepa-pending-approvals.md.superpowers/autopilot/state/pending-approvals.json.agents/skills/herald-autopilot/scripts/sync_pending_approvals.py.agents/skills/herald-autopilot/scripts/update_pending_approvals.pyDo not manually edit the state JSON or generated markdown queue except to repair tooling after a script failure.
Run from the repo root:
git status --short
git diff -- docs/superpowers/gepa-pending-approvals.md
If docs/superpowers/gepa-pending-approvals.md is already dirty, inspect the diff before running queue scripts. Proceed when it is only a generated queue refresh. Ask before overwriting if it contains unrelated hand-written edits.
Refresh the queue from published run reflections:
python3 .agents/skills/herald-autopilot/scripts/sync_pending_approvals.py --repo-root .
sed -n '1,140p' docs/superpowers/gepa-pending-approvals.md
Summarize the current pending items by title, queue key, why, approval prompt, and number of source runs.
Interpret user intent conservatively:
--all-pending.rejected, implemented, or pending only when the user explicitly asks for that status.Approve specific keys:
python3 .agents/skills/herald-autopilot/scripts/update_pending_approvals.py \
--repo-root . \
--status approved \
--key "<queue-key>" \
--note "Approved by user: <brief source phrase>"
Approve every pending item:
python3 .agents/skills/herald-autopilot/scripts/update_pending_approvals.py \
--repo-root . \
--status approved \
--all-pending \
--note "Approved by user: <brief source phrase>"
For other explicit decisions, replace approved with rejected, implemented, or pending. Keep notes short and factual.
Read back the updated queue:
sed -n '1,180p' docs/superpowers/gepa-pending-approvals.md
git diff -- docs/superpowers/gepa-pending-approvals.md
Confirm:
pending.superpowers state file and docs/superpowers/gepa-pending-approvals.mdDo not commit approval changes unless the user asks.
Report:
Herald beta release workflow for requests to cut, tag, publish, or repair a Herald release. Use when Codex should read RELEASE.md, review changes since the last tag, recommend the next v0.X.Y-beta.Z version, update What's New docs pages, create and push an annotated tag, monitor CI/CD, fix release blockers, move or supersede the tag when appropriate, and write release notes.
Use when you want one repo-local Herald workflow to take a single bug, feature, or workflow improvement from intake through planning, isolated worktree setup, implementation, impact-based verification, branch handoff, and GEPA-style run logging.
Herald-specific completion workflow for requests such as "finish-development", "nice" as a repo completion command, "wrap this branch", "fast merge", "fastpass", "quick merge", "fast mode", or "commit, merge to main, delete worktree". Use when a completed Herald task should be polished or fast-passed, committed, merged into main, and its feature worktree cleaned up.
Battle-test terminal UIs such as Herald via tmux and ttyd: reset states, compare visible hotkeys with actual behavior, capture ANSI/PNG evidence, run resize soaks, and verify terminal raster image previews.
Use when preparing a Herald beta release, checking release readiness, or running broad deterministic integration gates across TUI themes, inline images, SSH, and MCP before tagging.
Battle-test the Herald TUI — automated exploratory testing that captures screens, detects visual defects programmatically, and digs into anything suspicious. Simulates prolonged real usage.