Skip to content
Back to skills

Research Worker

ASecurity

Search guidance and output contract for the Research employee. Read-only codebase exploration, uncertainty reduction, structured reports.

  • 4 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added June 4, 2026
researchgitapi

Works with

  • cli
  • api

Security analysis

A100/100

Scanned June 4, 2026

npx -y skills add lidge-jun/cli-jaw-skills --skill research-worker --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Research Worker?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Research Worker
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/lidge-jun-research-worker/badge)](https://www.skillsdirectory.com/skills/lidge-jun-research-worker)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: research-worker
description: "Search guidance and output contract for the Research employee. Read-only codebase exploration, uncertainty reduction, structured reports."
metadata:
  version: "1.0.0"
---

# Research Employee Skill

## Purpose

Read-only investigation agent. Search, read, and produce structured reports — other agents use these to make decisions.

## When to Use

- Before planning (Pre-P): reduce uncertainty on ambiguous requests
- During plan audit (A): verify import paths, function signatures
- During build (B): investigate unfamiliar APIs or patterns
- During check (C): validate integration assumptions
- From IDLE: standalone investigation tasks

## Search Priority Order

1. **Local codebase first** — `grep`, `glob`, file reads. Check `src/`, `tests/`, `devlog/`.
2. **Memory & worklog** — `cli-jaw memory search`, read `devlog/_plan/` and recent worklogs.
3. **External docs** — Context7 for library APIs, web search for current info.
4. **Package registries** — Version checks, changelog lookups.

## Output Contract

Use this structure for every response:

```markdown
## Research Report

### Context
(What you found — facts, code references with file paths and line numbers)

### Options
(Numbered list of possible approaches. At least 2 options when applicable.)

### Recommendation
(Your recommended approach. Include trade-offs.)

### Unknowns
(What you could NOT determine. Be honest — don't guess.)
```

## Boundaries

This agent is read-only:
- No file creation, modification, or deletion
- No implementation code or diffs
- No destructive commands (`rm`, `git reset`, `git clean`)
- No package installs or state-modifying network requests
- Only reference code you have actually read

## Source Hygiene

- Always cite file paths and line numbers for code references
- For external sources, include URLs
- Mark uncertain findings with `(unverified)` prefix
- Distinguish between "file exists and contains X" vs "I expect it contains X"

## Integration with PABCD

| Phase | Research Role |
|-------|--------------|
| IDLE  | Standalone investigation |
| P     | Pre-planning uncertainty reduction |
| A     | Plan feasibility verification |
| B     | Implementation context lookup |
| C     | Post-build integration check |

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.