| name | discover |
| description | Find high-impact open source issues worth contributing to. Uses 5 search strategies (reactions, comments, help-wanted, good-first-issue, bugs) with configurable filters for stars, language, labels, and age. |
| argument-hint | [--stars=N] [--lang=LANG] [--strategy=NAME] |
/discover — Find OSS Contribution Targets
You are running the WondrAIWork discovery phase (Phase 01). Your job is to help the user find
high-impact open source issues worth contributing to.
Step 1: Understand the Search
Ask the user what they're looking for. If they provided arguments, use those. Otherwise ask:
- Language preference? (e.g., TypeScript, Rust, Go, Python — or "any")
- Minimum stars? (default: 1000)
- Strategy? (reactions, comments, help-wanted, good-first-issue, bugs — or "all")
- Any specific labels or keywords?
Step 2: Run Discovery Script
Execute the discovery script with the user's preferences:
bash ${CLAUDE_PLUGIN_ROOT}/scripts/discover.sh $ARGUMENTS
Available options:
--stars=N — Minimum repo stars (default: 1000)
--lang=LANG — Language filter
--label=LABEL — Additional label filter
--limit=N — Results per strategy (default: 5)
--max-age=DAYS — Maximum issue age (default: 90)
--strategy=NAME — Single strategy (reactions|comments|help-wanted|good-first-issue|bugs)
Step 3: Score the Results
Read the methodology scoring guide:
cat ${CLAUDE_PLUGIN_ROOT}/methodology/01-discover.md
For each promising issue from the results, apply the scoring matrix:
- Impact (weight 40%): reactions, user reports, core functionality affected
- Feasibility (weight 35%): scope clarity, tech familiarity, approach complexity
- Visibility (weight 25%): repo stars, org reputation
Formula: Total = (Impact x 0.40) + (Feasibility x 0.35) + (Visibility x 0.25)
| Score | Verdict |
|---|
| 4.0+ | Strong target — proceed to evaluate |
| 3.0-3.9 | Worth investigating |
| 2.0-2.9 | Marginal |
| <2.0 | Skip |
Step 4: Present Shortlist
Present a shortlist of 3-5 candidates in a table:
| # | Issue | Repo | Impact | Feasibility | Visibility | Total | Verdict |
|---|
For each candidate, include:
- Issue URL
- One-line summary of what the fix involves
- Key risks or concerns
Step 5: Suggest Next Steps
For the top candidate(s), suggest:
Use /evaluate <owner/repo> <issue-number> to do a deep evaluation before committing to the
contribution.
Important Notes
- Scores are preliminary at this stage — they get re-evaluated in Phase 02 (Evaluate) and Phase
03 (Understand)
- Check the "Red Flags to Avoid" section in the methodology for disqualifying signals
- Prefer issues that are: acknowledged by maintainers, less than 6 months old, with no competing PRs