القائمة
Validate HoP issue work, open a draft PR against main, run a second-agent review, and stop for user confirmation before merge.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
Pipeline simplificado: documento fonte → extração de conhecimento → padrões reutilizáveis → classificação contra repositório → canonical docs. Focado em gerar docs canônicos a partir de um único documento (currículo, análise, transcript). NÃO gera skills, exercícios ou integração profunda de currículo — apenas canonical docs + atualização de tags/aliases + system-of-record. Dispara com: 'canonicalize this document', 'extract canonical from', 'doc to canonical', 'create canonical docs from', 'documento para canonical', 'extrair canônicos de'.
Implementa uma camada de selecao de contexto model-agnostic e vendor-independent. Define um formato padronizado de contexto que qualquer modelo pode consumir, um Context Router que resolve queries de qualquer agente contra o grafo relacional e storage em tiers, e Vendor Adapters que traduzem o formato agnostico para o formato nativo de cada modelo. Transforma contexto no ativo mais duravel da organizacao — portavel entre modelos, sessoes, e provedores. Dispara com: 'neutral selection', 'model-agnostic context', 'vendor-independent context', 'context format standard', 'context router', 'multi-model context', 'portable context', 'cross-model context', 'camada neutra', 'contexto agnostico', 'formato de contexto', 'vendor adapter', 'neutral context layer', 'context portability', 'model migration context'.
Torna o retrieval de contexto budget-aware: cada candidato a retrieval e ranqueado por razao valor/custo antes de ser injetado no contexto. Implementa um Information Value Predictor que estima reducao de incerteza por candidato, um Token Cost Estimator que computa custo em tokens, e um Utility Feedback Loop que aprende com uso real — quais itens recuperados foram efetivamente referenciados pelo modelo. Previne o memory feedback loop onde o sistema construido para resolver o problema de memoria se torna o motor da degradacao. Contrapoe diretamente o Link 4 (inert memory feedback) do Agent Degradation Loop. Dispara com: 'selection-budgeted retrieval', 'budgeted retrieval', 'retrieval budget', 'cost-benefit retrieval', 'information value predictor', 'utility feedback retrieval', 'retrieval ranking', 'orcamento de retrieval', 'busca budgetada', 'recuperacao com orcamento', 'value-cost retrieval', 'retrieval utility feedback'.
Implementa armazenamento de contexto em tres tiers (hot/warm/cold) com promocao e democao baseadas em relevancia. Mantem o conjunto de trabalho ativo deliberadamente pequeno na hot tier (cache in-memory), move contexto recentemente relevante para warm tier (NVMe, baixa latencia), e arquiva historico completo na cold tier (object storage, alta latencia). O Tier Orchestrator executa transicoes baseadas em scores de relevancia e prefetch preditivo antes de cada passo de raciocinio. Previne context rot (cada token acumulado degrada qualidade do passo seguinte) e separa preocupacao de armazenamento de preocupacao de selecao. Dispara com: 'tiered context', 'tier storage', 'context tiers', 'hot warm cold', 'tiered storage', 'context promotion', 'context demotion', 'tier orchestrator', 'promocao de contexto', 'democao de contexto', 'armazenamento em tiers', 'tiered context storage', 'context tier management'.
Operacionaliza um curriculo de autonomia para agentes: controla a proporcao entre rollouts supervisionados (professor/humano) e rollouts auto-gerados (agente) com um parametro lambda, gates de prontidao por classe de tarefa, e uma progressao explicita observe→assist→own. Aplica importance sampling para corrigir a distribuicao mista. Previne cold-start collapse (rollouts puros do agente antes de aprender recuperacao) e autonomy stagnation (agente nunca pratica recuperacao autonoma). Usar ao implantar um novo agente em producao, ao fazer transition de fluxo manual para agentico, ao calibrar o grau de supervisao de um agente existente, ou quando o agente apresenta comportamento fragil em cenarios nao-supervisionados. Dispara com: 'autonomy curriculum', 'curriculo de autonomia', 'lambda schedule', 'teacher mixing', 'rollout sampling', 'autonomy dial', 'autonomy progression', 'observe assist own', 'readiness gate', 'student rollout', 'teacher-student mix', 'gradual autonomy', 'agente semi-supervisionado', 'autonom
Separa o sinal de melhoria do agente em dois componentes independentes: magnitude (o quanto o modelo acredita que uma mudanca importa, extraida de self-distillation delta, log-ratio, atencao, ou confianca interna) e direcao (para onde a mudanca deve ir, determinada por verificador externo, testes, ou revisao humana). Combina magnitude × direcao em um plano de correcao ponderado: gaste esforco de correcao onde o agente tem conviccao FORTE e o verificador confirma a direcao; reduza ou escale quando a direcao e incerta. Previne information leakage (agente aprende a imitar formato sem substancia) e overconfidence collapse (self-distillation puro sem verificacao externa). Usar ao projetar loops de melhoria de agente, ao implementar self-distillation com verificacao, ao calibrar feedback de producao, ou quando o agente produz outputs com formato correto mas conteudo errado. Dispara com: 'magnitude direction', 'verifier split', 'trust but verify', 'confidence direction', 'correction weight', 'self-distillation verif
| name | issue-review |
| description | Validate HoP issue work, open a draft PR against main, run a second-agent review, and stop for user confirmation before merge. |
| license | MIT |
| compatibility | opencode |
| metadata | {"audience":"all-agents","workflow":"github","priority":"high"} |
I sit between implementation and merge. I validate the worktree with HoP's real npm gates, create a draft PR targeting main, run a second-agent review on the diff, surface findings, and stop. Nothing merges until the user explicitly confirms.
Load this skill when:
You need:
Nissue/<N>-<slug>.worktrees/<N>-<slug>If any are unclear, ask the user or inspect with git worktree list and gh issue view.
Run /compact before CI and PR creation on non-trivial work. Enter review with a clean context focused on the diff and validation output.
Always run the relevant gates from package.json; do not invent commands.
Core review gates:
WORKTREE=".worktrees/N-SLUG"
npm --prefix "$WORKTREE" run test:regression:mock
npm --prefix "$WORKTREE" run ops:preflight
Add surface-specific gates when relevant:
Eval-sensitive changes include prompt, model, tool, context, memory, rubric, evaluator policy, agent-loop, sampling, corpus, or rollout-threshold changes. For those PRs, select validation from the eval tier registry documented in curriculum/07-implementation-guides/06-harness-evolution-playbook.md and save the summary for the PR body's Eval impact section.
npm --prefix "$WORKTREE" run lint
npm --prefix "$WORKTREE" run test:unit
npm --prefix "$WORKTREE" run test:integration
npm --prefix "$WORKTREE" run dashboard:test
npm --prefix "$WORKTREE" run dashboard:build
npm --prefix "$WORKTREE" run test:fixture-parity
npm --prefix "$WORKTREE" run evidence:verify
npm --prefix "$WORKTREE" run test:ci-gates
npm --prefix "$WORKTREE" run ops:verify-branch-protection
npm run smoke:live requires live runtime/auth context and should be run only when the issue touches live runtime behavior and the user has approved the external interaction.
If a validation step fails, stop. Show the relevant error and fix before creating or updating the PR.
Save validation output summaries for the PR body.
For eval-sensitive PRs, also capture:
case_id or trace_id examples with expected behavior and current decision.Before opening the PR, verify:
.runtime/ and artifacts/ changes, if any, respect HOP_TENANT_ID tenant isolation.Eval impact section in the PR body, not only generic test output..env.example and relevant docs.DESIGN.md.docs/canonical/, docs/guides/, docs/evidence/, or ADRs as appropriate.The PR targets main and should follow .github/PULL_REQUEST_TEMPLATE.md.
ISSUE_TITLE=$(gh issue view N --json title --jq '.title')
BRANCH="issue/N-slug"
WORKTREE=".worktrees/N-slug"
DIFF_STAT=$(git -C "$WORKTREE" diff --stat origin/main...HEAD)
COMMITS=$(git -C "$WORKTREE" log origin/main...HEAD --oneline)
PR_BODY=$(cat <<'PR_EOF'
## Resumo
<1-3 frases explicando o que mudou e por que.>
Closes #N
## Mudanças
- <Mudança principal>
## Testes
- [ ] `npm run test:regression:mock` passou
- [ ] `npm run ops:preflight` passou
- [ ] Nenhum teste novo quebrou
---
## Eval impact
N/A
## Crossroad-file impact
N/A
PR_EOF
)
git -C "$WORKTREE" push -u origin "$BRANCH"
gh pr create \
--title "$ISSUE_TITLE" \
--body "$PR_BODY" \
--base main \
--head "$BRANCH" \
--draft
Fill the PR template accurately. If the diff touches prompt, model, tool, context, memory, rubric, or agent-loop behavior, replace Eval impact N/A with the eval report summary from Step 2. If any crossroad file changed, replace N/A with the full crossroad section: affected files, change type, migration/consumer impact, regression proof, mock parity if relevant, and code-owner approval expectation.
WORKTREE=".worktrees/N-SLUG"
git -C "$WORKTREE" diff origin/main...HEAD
git -C "$WORKTREE" diff --stat origin/main...HEAD
git -C "$WORKTREE" log origin/main...HEAD --oneline
gh issue view N --json title,body --jq '"# " + .title + "\n\n" + .body'
Delegate a review subagent with this scope:
TASK: Review the HoP diff for issue #N.
Review for:
- Correctness against issue scope and acceptance criteria
- Minimal change; no unrelated cleanup
- Tests and validation coverage for the changed surface
- Tenant isolation for `.runtime/` and `artifacts/`
- Crossroad file policy and PR template completeness
- Supabase client/mock parity if persistence files changed
- pino logging and redaction safety
- Config/env var documentation (`.env.example` and docs)
- Dashboard compliance with `DESIGN.md` if UI changed
- Documentation precedence and ADR conflicts
- Eval-sensitive PR evidence: tier registry selection, baseline/candidate versions, quality/latency/cost delta, threshold result, skipped-tier waiver, and failure examples
- Security: no secrets, no type suppressions, no unsafe external side effects
Report findings as BLOCKING or ADVISORY. Do not rewrite code.
Collect findings before proceeding.
Run /compact after receiving review findings and before presenting the final review summary to the user.
Report:
## Review Complete - Issue #N: <title>
What we had to do:
<issue summary>
What changed:
<commit list and diff stat>
Validation:
- <commands run and outcomes>
Second-agent review:
- BLOCKING: <items or none>
- ADVISORY: <items or none>
PR: <draft PR URL>
To proceed, confirm "ship it" and I will load issue-finish.
To fix findings, update the worktree and rerun issue-review.
Do not merge. Wait for explicit user confirmation.
main.npm --prefix .worktrees/N-slug run test:regression:mock
npm --prefix .worktrees/N-slug run ops:preflight
git -C .worktrees/N-slug push -u origin issue/N-slug
gh pr create --base main --head issue/N-slug --draft
git -C .worktrees/N-slug diff origin/main...HEAD