| name | quick-research |
| description | Fast fact check via GREP (~25 seconds). Use for simple lookups where you need a sourced answer but don't need a deep investigation — API endpoint verification, version checks, "what's the current X" questions, quick pre-code sanity checks. Prefer this over /research when a 25-second answer beats a 2-minute answer. Use /research instead if you need comprehensive coverage, or /ultra-research for full investigations. |
Quick Research (effort=low)
Fastest GREP tier. ~25 seconds end-to-end. Single command, single blocking call, returns the report.
When to use this skill vs the others
| Skill | Time | Use when |
|---|
/quick-research (you are here) | ~25s | Quick fact, version check, API endpoint lookup, "what's new in X" |
/research | ~5 min | Most research tasks, comprehensive docs, library investigation |
/ultra-research | up to 1 hour | Security audits, legal research, full ecosystem surveys |
Rule of thumb: if you'd be happy with a single well-sourced paragraph, use this. If you need structured coverage of multiple angles, use /research. If you need exhaustive multi-source synthesis, use /ultra-research.
Prerequisites
The user must be authenticated. Before running any API call, check auth status:
node "$SCRIPTS_DIR/auth.js" status
If the result shows "authenticated": false, automatically invoke /grep-login — don't just tell the user to do it. Run the login flow, then continue with the research once authenticated.
Resolve the script path
The skill dir is usually symlinked, so always resolve:
SCRIPTS_DIR="$(dirname "$(dirname "$(dirname "$(readlink -f "${CLAUDE_SKILL_DIR}/SKILL.md")")")")/scripts"
Run it — use Monitor (background), NOT blocking Bash
IMPORTANT: Always run research as a background Monitor task, never as a blocking Bash call. Even though quick-research only takes ~25 seconds, Monitor keeps the agent responsive and streams live updates.
SCRIPTS_DIR="$(dirname "$(dirname "$(dirname "$(readlink -f "${CLAUDE_SKILL_DIR}/SKILL.md")")")")/scripts"
node "$SCRIPTS_DIR/grep-api.js" run "$ARGUMENTS" --effort=low --max-wait=60 2>&1
Run this with Monitor (timeout_ms: 80000, persistent: false). With 2>&1, status updates and the final report both stream as events.
CRITICAL: Always deliver results
When the Monitor notification arrives saying the task completed, you MUST:
- Read the output file from the task notification
- Extract and present the research report to the user
- Never silently drop a completed research job
Fallback: blocking Bash (only if Monitor is unavailable)
Only use blocking Bash if the Monitor tool is genuinely not available:
node "$SCRIPTS_DIR/grep-api.js" run "$ARGUMENTS" --effort=low --max-wait=60
Set the Bash tool timeout to at least 80000 (80 seconds).
Presenting results
The output is typically short — a direct answer with 1-3 citations. Present it clearly:
- Lead with the answer
- Preserve citations
- If the query has nuance the ultra_fast tier missed, suggest running
/research for deeper coverage
If the job times out
Exit code 2 means the server is still working. The JSON payload will include a job_id. Use /grep-status with the job ID to resume, or wait and rerun.
Anti-patterns
- Do NOT use
/quick-research for complex investigations. If the question has multiple sub-questions or needs cross-referencing, use /research instead.
- Do NOT invoke with the default 120s bash timeout without the
--max-wait=60 cap — Node needs to exit before the bash tool kills it.
