re
starsforks
How to install re?
dgk-dev/dgk-gpt/re
$npx -y skills add dgk-dev/dgk-gpt --skill reExplicit extra-research mode. Use only when the user says /re or clearly asks for a research-heavy pass before coding. This skill should bias Codex toward more source-checking and justification without replacing its normal orchestration.
SKILL.md
| 1 | # Re |
| 2 | |
| 3 | Use this as a light override, not a rigid workflow. Keep Codex's default orchestration and only add extra diligence where it matters. |
| 4 | |
| 5 | ## Defaults |
| 6 | |
| 7 | - Inspect the local codebase first. |
| 8 | - Do research in 3 passes when the task is genuinely research-heavy: |
| 9 | 1. Plan 3-6 sub-questions. |
| 10 | 2. Retrieve sources for each sub-question and follow 1-2 second-order leads when useful. |
| 11 | 3. Synthesize only after resolving conflicts and gaps that could change the conclusion. |
| 12 | - Use Context7 first for version-sensitive library or framework behavior. |
| 13 | - Use native `web` first for current information. |
| 14 | - Use Jina only when native `web` is not enough for long pages, PDFs, or parallel page reads. |
| 15 | - Only cite sources retrieved in the current workflow, and label any inference that is not directly supported by those sources. |
| 16 | - If sources conflict, state the conflict explicitly instead of averaging them into one answer. |
| 17 | - If a search result is empty, partial, or suspiciously narrow, retry with a broader query, alternate wording, or a second source before concluding there is no answer. |
| 18 | - Compare options only when there is a real tradeoff. |
| 19 | - Implement once the decision is clear. |
| 20 | - Run the smallest relevant verification commands before finishing. |
| 21 | |
| 22 | ## Avoid |
| 23 | |
| 24 | - Do not force phases. |
| 25 | - Do not produce long research reports unless the user asked for them or the task materially benefits. |
| 26 | - Do not ask extra questions if local context is already enough. |
| 27 | - Do not delegate work just because this skill is active. |
| 28 | |
| 29 | ## Large Tasks |
| 30 | |
| 31 | For long tasks, optional scratch notes can live in `/tmp/re-research/<slug>/`. |
| 32 | |
| 33 | If useful, keep a short `decisions.md` with: |
| 34 | - key findings |
| 35 | - rejected options |
| 36 | - remaining risks |
| 37 | - pending verification |
| 38 | |
| 39 | ## Finish |
| 40 | |
| 41 | Keep the final response compressed: |
| 42 | - decision |
| 43 | - source-backed justification |
| 44 | - key sources or citations when they materially matter |
| 45 | - code or config changed |
| 46 | - tests run |
| 47 | - remaining unknowns |
People who liked this also viewed
6 picksType
- find-skillsHelps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skSkillsJul 2026
- grill-meA relentless interview to sharpen a plan or design.SkillsJul 2026
- vercel-react-best-practicesVercel's official collection of agent skillsSkillsJul 2026
- improve-codebase-architectureScan a codebase for deepening opportunities, present them as a visual HTML report, then grill through whichever one you pick.SkillsJul 2026
- setup-matt-pocock-skillsConfigure this repo for the engineering skills — set up its issue tracker, triage label vocabulary, and domain doc layout. Run once before first use of the other engineering skills.SkillsJul 2026
- triageMove issues and external PRs through a state machine of triage roles — categorise, verify, grill if needed, and write agent-ready briefs.SkillsJul 2026
