jules-recovery-protocol
Protocolo de recuperación para Jules cuando falla una sesión.
Skill by iberi22
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/iberi22-jules-recovery-protocol
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: agent-building
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/jules-recovery-protocol
What's inside
- Skill name
jules-recovery-protocol- Description
- Protocolo de recuperación para Jules cuando falla una sesión. Incluye prompt optimizado en 3 fases: Diagnóstico → Investigación Enfocada → Implementación+PR. Versión validada con Ola 8 Xavier.
- License
- MIT
- Instructions
- 93 lines in SKILL.md · view on GitHub
Files (1)
- SKILL.md
SKILL.md (first lines)
# Jules Recovery Protocol — Prompt Optimizado
> Cuando Jules falla, el prompt de recovery es crítico. Un mal prompt = Jules vuelve a fallar.
> Este skill contiene el prompt probado y las reglas para aplicarlo.
## 🚨 Diagnóstico Rápido
Antes de recovery, clasificar la falla:
| Síntoma | Causa | Acción |
|---------|-------|--------|
| "unexpected error" | Task creada pero falló en ejecución | Comentar recovery prompt en el issue |
| "failed to create a task" | Rate limit / server error | Remover label jules → re-aplicar |
| "Completed" sin cambios | No pudo pushear o task era research | Verificar git diff, relanzar |
| PR con errores de CI | Tests fallaron | Fix manual o relanzar con más contexto |
## Prompt de Recovery (PROBADO)
**CUÁNDO USARLO:** Cuando Jules comenta "I apologize, but I encountered an unexpected error"
**CÓMO USARLO:** Comentar en el issue como respuesta al mensaje de error de Jules.
### Prompt Mejorado v2
```
## 🔄 Recovery — Research + Implementation Protocol
### Phase 1: Diagnosis (2 min)
1. Read the ENTIRE issue body — every section matters
2. Identify the EXACT failure: compile error? test failure? missing file? wrong path?
3. Read the last error message from your previous attempt (if available)
### Phase 2: Focused Web Research (5 min)
Only research what's BLOCKING you:
- `search: "[exact error message or library name] 2026 rust"`
- `search: "[alternative approach for the specific blocking issue]"`
- `search: "[crate documentation or reference implementation]"`
Do NOT research concepts already working. Only the blocker.
### Phase 3: Implementation & Validation
1. Write MINIMAL code solving the issue — follow "Files to Modify" section
2. For EACH Acceptance Criteria in the issue body, verify:
- `grep -c "[expected pattern]" [file]` matches
- `cargo check --workspace` → 0 errors
- `cargo test -p xavier --lib [module]` → 0 failures
3. If stuck after 3 attempts: post the EXACT failing command + output
### Hard Rules
- ONLY modify files listed in "Files to Modify" or "DO NOT touch" allows
- NO Cargo.toml changes unless the issue body explicitly requires it
- NO new files unless "NEW" is specified in the issue
- If you can't make it compile, submit a draft PR anyway with the closest working version
### Recovery Output
After successful implementation:
1. Branch: `fix/recovery-{issue-number}-{short-description}`
2. Commit: `feat: {feature description} (#{issue})`
3. Push + Create PR
4. Comment on issue: "Fixed in PR #{number}"
```
## Comparativa: Prompt Antiguo vs Prompt Mejorado
| Aspecto | Prompt Antiguo | Prompt Mejorado v2 |
|---------|---------------|-------------------|
| Estructura | 1 oración larga | 3 fases claras + hard rules |
| Diagnóstico | ❌ No pide identificar la falla | ✅ Fase 1: Diagnosis |
| Investigación | "investigación en la web" (vaga) | ✅ Queries específicas para el bloqueo |
| Referencia al issue | ❌ No menciona el body | ✅ "Read the ENTIRE issue body" |
| Acceptance Criteria | ❌ No las menciona | ✅ Verificar contra cada AC |
| Output esperado | ❌ No especifica formato | ✅ Branch → Commit → PR |
| Fallback | ❌ Ninguno | ✅ "Post exact failing command" |
| Límites | ❌ No protege archivos | ✅ Hard Rules con "DO NOT touch" |
| Typo | "verificacio" | ✅ Corregido |
## Reglas de Aplicación
1. **Siempre leer el último comentario de Jules** antes de aplicar recovery
2. **Si Jules ya empezó a implementar algo parcial**, preservar su progreso
3. **Para issues con PR parcial**, comentar en el PR, no en el issue
4. **Para issues sin task creada** (#992, #993): remover label jules → esperar 5s → re-aplicar
5. **Si recovery falla 2 veces seguidas**, el issue necesita revisión humana del body
## Post-Recovery Checklist
- [ ] Jules reconoció el recovery ("on it" comment)
- [ ] Task ID creada (no "failed to create task")
- [ ] PR creado en < 10 min
- [ ] PR pasa CI (cargo check + test)
- [ ] Acceptance Criteria cumplidos
- [ ] features.json actualizadoFrom iberi22/skills, MIT.
More from iberi22/skills
- 0.570score
agent-issue-delegation
iberi22
Crear GitHub Issues de calidad profesional para delegar a agentes autónomos (Jules, Claude Code, Codex, OpenCode).
SkillAFreshness: AClaude Code +9 - 0.567score
linux-performance-profiles
iberi22
Switchable Linux system performance profiles (gaming, sen/headless work, default) via sysfs/sysctl/Hyprland.
SkillAFreshness: AClaude Code +9 - 0.566score
cron-fleet-maintenance
iberi22
Use when auditing or repairing a fleet of cron jobs.
SkillAFreshness: AClaude Code +9 - 0.565score
issue-triage
iberi22
Use when moving issues or external PRs through the issue tracker's triage state machine (needs-triage, ready-for-agent, wontfix).
SkillAFreshness: AClaude Code +9 - 0.565score
bug-triage
iberi22
Use when diagnosing a bug in code.
SkillAFreshness: AClaude Code +9 - 0.561score
swal-theme-creator
iberi22
Allows AI agents (Hermes, Antigravity, Jules) to generate, validate, and switch visual themes in SWAL Desktop using declarative JSON tokens.
SkillAFreshness: AClaude Code +9 - 0.560score
rclone-cloud-backup
iberi22
Use when auditing a running rclone cloud sync job.
SkillAFreshness: AClaude Code +9 - 0.560score
cpanel-node-ops
iberi22
Use when deploying Node backends on cPanel hosting.
SkillAFreshness: AClaude Code +9 - 0.557score
hermes-gateway-troubleshooting
iberi22
Debug Hermes Gateway connectivity and platform integration issues — Telegram, Discord, WhatsApp.
SkillAFreshness: AClaude Code +9 - 0.556score
background-transfer-monitor
iberi22
Use when running long background uploads: track progress.
SkillAFreshness: AClaude Code +9 - 0.521score
caveman
iberi22
Ultra-compressed communication mode para coding agents.
SkillBFreshness: BClaude Code +9 - 0.521score
claude-design-principles
iberi22
Principios de diseño UI/UX distillados del system prompt de Claude Design.
SkillBFreshness: BClaude Code +9
Signals
- Stars
- 0
- Forks
- 0
- Open issues
- 3
- Last commit (this folder)
- Oct 8, 2026 (21 days ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/iberi22/skills
- Website
- github.com/iberi22/skills
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from iberi22
- 0.570score
agent-issue-delegation
iberi22
Crear GitHub Issues de calidad profesional para delegar a agentes autónomos (Jules, Claude Code, Codex, OpenCode).
SkillAFreshness: AClaude Code +9 - 0.567score
linux-performance-profiles
iberi22
Switchable Linux system performance profiles (gaming, sen/headless work, default) via sysfs/sysctl/Hyprland.
SkillAFreshness: AClaude Code +9 - 0.566score
cron-fleet-maintenance
iberi22
Use when auditing or repairing a fleet of cron jobs.
SkillAFreshness: AClaude Code +9 - 0.565score
issue-triage
iberi22
Use when moving issues or external PRs through the issue tracker's triage state machine (needs-triage, ready-for-agent, wontfix).
SkillAFreshness: AClaude Code +9 - 0.565score
bug-triage
iberi22
Use when diagnosing a bug in code.
SkillAFreshness: AClaude Code +9 - 0.561score
swal-theme-creator
iberi22
Allows AI agents (Hermes, Antigravity, Jules) to generate, validate, and switch visual themes in SWAL Desktop using declarative JSON tokens.
SkillAFreshness: AClaude Code +9
Cite this page
jules-recovery-protocol skill. Agent Mart. https://agentmart.dev/skills/iberi22-jules-recovery-protocol (last updated 2026-10-24)