
Investigate First
Makes the agent diagnose unclear bugs with evidence before it edits any code.
- Repository
- JuliusBrussee/caveman
- Stars
- 116,835
- Licence
- Apache-2.0
- Files
- 2 files
What it does
When a bug has an unknown cause, behaves intermittently, or a system has slowed down, this skill tells the agent to gather evidence before changing anything. The agent separates what it actually observed from what it guesses, traces inputs, state changes and error output, and ranks possible causes by how well the evidence supports them. It only edits code once one cause clearly explains the evidence, then reports the cause and the proof.
Great for
- Developers stuck on a bug whose cause is not obvious
- Teams investigating intermittent failures or performance regressions
- Anyone who wants an explanation with evidence before a fix is applied
Skip it if
- Simple, well-understood fixes where investigation adds little value, since the skill slows the agent down
- Tasks where the user wants the agent to implement a change right away, because the skill says not to fix unless implementation is authorized
Try asking
- “Our checkout API sometimes returns 500 errors under load. Find out why before changing anything.”
- “Page load time doubled after last week's release. Investigate the cause and show the evidence.”
- “This test passes locally but fails randomly in CI. Diagnose it first.”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 2 original files, the Apache-2.0 licence and an attribution note.
investigate-first.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip investigate-first.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip investigate-first.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- State the symptom clearly and include error messages or logs so the agent has evidence to work from.
- Say explicitly whether you want a fix or only a diagnosis, since the skill defaults to diagnosis only.
- Ask the agent to show its ranked hypotheses so you can check its reasoning.
What's inside
- agents/openai.yaml
- SKILL.md
Redistributed unmodified from JuliusBrussee/caveman under the Apache-2.0 licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.