All skills
simota avatar

/latch

@16a3f28
by shingo imotasimota/agent-skills85 stars
15

Proposing, configuring, debugging, and maintaining Claude Code hooks (PreToolUse/PostToolUse/Stop and other lifecycle events). Use for workflow automation or quality gates via hooks.

Use this Skill: https://skilld.dev/gh/simota/agent-skills/latch

This session only. Nothing lands on disk.

referencenexus-integration.md

≈653 tokens on demand. Your agent reads this file only when SKILL.md points to it.

Nexus Integration

Purpose: Read this when Latch runs under Nexus AUTORUN, when the input contains ## NEXUS_ROUTING, or when you need _STEP_COMPLETE: / ## NEXUS_HANDOFF templates.

Contents

AUTORUN Agent Context

When Nexus invokes Latch in AUTORUN mode:

_AGENT_CONTEXT:
  Role: Latch
  Task: [Specific hook task]
  Input:
    request: [User request]
    current_hooks: [Current hook summary]
    project_type: [Project type]
  Constraints:
    - Change only the `hooks` section in settings.json
    - Create a backup before modification
    - Remind the user to restart the session
  Expected_output:
    - Settings changes
    - Scripts created
    - Verification result

AUTORUN Step Completion

After Latch finishes, append:

_STEP_COMPLETE:
  Agent: Latch
  Status: SUCCESS|PARTIAL|BLOCKED|FAILED
  Output:
    hooks_configured: [Configured hooks]
    scripts_created: [Created scripts]
    settings_modified: true|false
    backup_created: true|false
    verification: [Verification summary]
  Risks:
    - [Risk item if any]
  Reminders:
    - Session restart required
  Next: [Next agent]|DONE

NEXUS_HANDOFF Format

When returning results to Nexus hub, use ## NEXUS_HANDOFF and preserve the broader field contract defined in SKILL.md.

## NEXUS_HANDOFF

### Step: Latch
**Agent:** Latch
**Summary:** [One-line summary]

### Key findings
- [Finding]

### Artifacts
- **settings.json:** [Changed content]
- **Scripts:** [Created or updated scripts]
- **Backup:** [Backup information]

### Verification
- `/hooks`: [Result]
- `claude --debug`: [Result]
- Script test: [Result]

### Risks
- [Risk or follow-up]
- Session restart required

### Open questions
- [Question if any]

### Pending Confirmations (Trigger/Question/Options/Recommended)
- [Confirmation item if any]

### User Confirmations
- [Accepted confirmation if any]

### Next
**Suggested next agent:** [Next agent]
**Next action:** [Next action]

Nexus Routing Integration

Task Type Chain Use when
HOOKS Latch Configure or debug Claude Code hooks
HOOKS/security Sentinel -> Latch Turn security requirements into hooks
HOOKS/quality Latch -> Radar Pair quality-gate hooks with test verification
INFRA/hooks Latch -> Gear Connect hooks with infrastructure or CI/CD work

Source: SKILL.md on GitHub

1 warning5mo5 checks · Risk SAFE
  • Gen Agent Trust Hub5mo

    The skill is a specialized tool for managing Claude Code hooks. While it provides significant security benefits through pre-built safety recipes for secret detection and command guarding, it inherently handles the generation and configuration of executable shell scripts. This creates a surface for indirect prompt injection if the agent is directed to process malicious project-level configuration files.

  • Socket5mo

    No alerts

  • Snyk5mo

    Risk: LOW · No issues

  • Runlayer6mo

    3/5 files flagged

  • ZeroLeaks5mo

    Score: 93/100 · 2 sections analyzed

Signed by skilld at 16a3f28. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 2 days ago.

Activeupdated last month

README badge

README badge for simota/agent-skills/latch