lucy-agent

Files (74)

• ci.yml · .github/workflows/
• .gitignore
• CHANGELOG.md
• LICENSE
• PR_REVIEW_INSTRUCTIONS.md
• README.md
• RELEASE_NOTES.md
• SKILL.md
• SPEC.md
• clawhub-skills.txt
• agent-fragment.json5 · config/
• STANDARDS.md · docs/
• install.sh
• check-content-boundaries.sh · scripts/
• common.sh · scripts/
• export-full-clone.sh · scripts/
• install-pre-commit-hook.sh · scripts/
• design.md · sdd/lucy-ai/english-documentation-language-matching/
• explore.md · sdd/lucy-ai/english-documentation-language-matching/
• spec.md · sdd/lucy-ai/english-documentation-language-matching/
• state.json · sdd/lucy-ai/english-documentation-language-matching/
• tasks.md · sdd/lucy-ai/english-documentation-language-matching/
• verify.md · sdd/lucy-ai/english-documentation-language-matching/
• apply.md · sdd/lucy-ai/full-clone-migration-2026-07-31/
• design.md · sdd/lucy-ai/full-clone-migration-2026-07-31/
• spec.md · sdd/lucy-ai/full-clone-migration-2026-07-31/
• state.json · sdd/lucy-ai/full-clone-migration-2026-07-31/
• tasks.md · sdd/lucy-ai/full-clone-migration-2026-07-31/
• verify.md · sdd/lucy-ai/full-clone-migration-2026-07-31/
• apply.md · sdd/lucy-ai/release-configuracion-agnostica-enforcement/
• design.md · sdd/lucy-ai/release-configuracion-agnostica-enforcement/
• explore.md · sdd/lucy-ai/release-configuracion-agnostica-enforcement/
• spec.md · sdd/lucy-ai/release-configuracion-agnostica-enforcement/
• state.json · sdd/lucy-ai/release-configuracion-agnostica-enforcement/
• tasks.md · sdd/lucy-ai/release-configuracion-agnostica-enforcement/
• verification.md · sdd/lucy-ai/release-configuracion-agnostica-enforcement/
• design.md · sdd/lucy-ai/release-live-workspace-sync-2026-07-29/
• pr-body.md · sdd/lucy-ai/release-live-workspace-sync-2026-07-29/
• spec.md · sdd/lucy-ai/release-live-workspace-sync-2026-07-29/
• state.json · sdd/lucy-ai/release-live-workspace-sync-2026-07-29/
• tasks.md · sdd/lucy-ai/release-live-workspace-sync-2026-07-29/
• verify.md · sdd/lucy-ai/release-live-workspace-sync-2026-07-29/
• orchestrator-flow.md · sdd/
• task-string-format.md · sdd/
• apply.md.in · sdd/templates/
• design.md.in · sdd/templates/
• explore.md.in · sdd/templates/
• spec.md.in · sdd/templates/
• standards.md.in · sdd/templates/
• state.json.in · sdd/templates/
• tasks.md.in · sdd/templates/
• verify.md.in · sdd/templates/
• validation-rules.md · sdd/
• uninstall.sh
• update.sh
• verify.sh
• AGENTS.md · workspace/
• HEARTBEAT.md · workspace/
• IDENTITY.md · workspace/
• MEMORY.md · workspace/
• …and 14 more

SKILL.md (first lines)

## What is lucy-agent?

lucy-agent is the installable ecosystem for Lucy-ai — a fully configured AI colleague
with personality, skills, conventions, and gateway config fragment.

**Repository:** https://github.com/camiloandresgtruniandes/lucy-agent

## Installation

### One-command install (fresh machine)

```bash
curl -fsSL https://raw.githubusercontent.com/camiloandresgtruniandes/lucy-agent/main/install.sh | bash
```

### Flags

- `--skip-clawhub` — Skip ClawHub skill installation
- `--skip-workspace` — Skip workspace file seeding
- `--force` — Overwrite conflicting files without prompting
- `--dry-run` — Show what would be done without making changes

### Post-install (manual)

1. Link the config fragment in `openclaw.json`:
   ```json5
   { $include: "./lucy-agent/config/agent-fragment.json5" }
   ```
2. Configure channel tokens and API keys in `openclaw.json`
3. Restart OpenClaw: `openclaw gateway restart`

## Updating

```bash
cd ~/.openclaw/lucy-agent && ./update.sh
# Or pin to a specific version:
./update.sh --tag v1.0.0
```

## Verification

```bash
cd ~/.openclaw/lucy-agent && ./verify.sh
```

## What's included

- **Bundled skills:** sdd, github-pr, pr-review, csharp-dotnet, tailwind-4, typescript, skill-creator
- **ClawHub skills:** weather, browser-automation, acp-router (auto-installed)
- **Workspace seed:** SOUL.md, IDENTITY.md, AGENTS.md, USER.md, TOOLS.md, HEARTBEAT.md
- **Config fragment:** agent defaults (skills, model, thinking level)

## Conflict handling

lucy-agent is **idempotent and careful**:
- If a workspace file was modified by the user, it prompts before overwriting
- If a skill was modified by the user, it prompts before overwriting
- Re-running install.sh on an existing installation does a `git pull` (not a fresh clone)

## Security

- No secrets are ever committed (API keys, tokens, credentials)
- The install script is auditable — curl it and read before running
- Config fragment contains only public configuration (skill names, model names, flags)

## Troubleshooting

### "OpenClaw CLI not found"
Install OpenClaw first: `curl -fsSL https://openclaw.ai/install.sh | bash`

### verify.sh fails
Run with verbose output: `bash -x verify.sh`
Check that all skills in `skills/` have a `SKILL.md` file

### git pull fails on existing installation
```bash
cd ~/.openclaw/lucy-agent
git status  # check for uncommitted changes
git stash   # or manually handle conflicts
git pull
```

From CamiloAndresGTRUniandes/lucy-ai, MIT.

More from CamiloAndresGTRUniandes/lucy-ai