manastalukdar/ai-devstudio/research

Research

Spin up a background agent to do the research, so you keep working while it reads.

Background agent job

1. Investigate the question against primary sources (official docs, source code, specs, first-party APIs) — not secondary write-ups. Follow every claim back to the source that owns it.
2. Write the findings to a single Markdown file, citing each claim's source.
3. Save it where the repo already keeps such notes; match the existing convention. If there is none, put it somewhere sensible and say where.

Usage

/research "how does Postgres LISTEN/NOTIFY work under high connection load?"
/research "what are the rate limits for the GitHub API?"
/research "what changed in React 19 that affects useEffect cleanup?"

Guidelines

• Primary sources over summaries: official docs, RFCs, changelog entries, source code comments.
• Cite everything: every claim in the output file links to the source that makes it.
• No hallucination: if a primary source can't be found for a claim, say so explicitly in the file rather than asserting the claim.
• The output file is a primary source pointer, not a conclusion — it feeds into /grill-with-docs or /to-spec, where decisions are made.

Output location

Check for existing research/notes conventions in this order:

1. docs/research/
2. .scratch/research/
3. research/
4. Root of repo (last resort; state the path clearly)

Save as <slug>.md where slug is derived from the question.

Token Optimization

Expected range: 50–200 tokens (orchestrator only; research work runs in background subagent)

Patterns used: Subagent isolation keeps orchestrator context minimal; background agent runs while main session continues

Ported from [mattpocock/skills](https://github.com/mattpocock/skills) with attribution.