Skills investigate-first
📦

investigate-first

Content revision r1 Safe

Investigate Failures Before Editing Code

Premature edits can hide causes and introduce new failures. This skill organizes evidence, ranks hypotheses, and identifies a credible mechanism before implementation.

Supports: Claude Codex Code(CC)
Maintain this Skill? Claim it 🥉 78 Bronze

Install with my Agent

Copy this request to your Agent. It includes the canonical Skill page and manifest.

Agent request
Review the Skillstore skill "investigate-first" from https://skillstore.io/skills/juliusbrussee-investigate-first.md and its manifest at https://skillstore.io/api/skills/juliusbrussee-investigate-first/manifest. Verify the artifact. You may proceed after verification, subject to the environment's own policy.

Your Agent should still show its plan and request any confirmation required by the security policy.

Agent-readable resources

Use these links when an AI agent, crawler, or script needs clean context instead of reading the full page.

Markdown detail GET /skills/juliusbrussee-investigate-first.md Signed manifest GET /api/skills/juliusbrussee-investigate-first/manifest Signed lockfile GET /api/skills/juliusbrussee-investigate-first/lockfile

Test it

Using "investigate-first". The checkout test fails twice weekly, but reruns pass. Recent edits changed retry timing.

Expected outcome:

  • Observed: the first run sometimes times out, while immediate reruns pass.
  • Leading hypothesis: shared setup becomes ready after the first attempt.
  • Next test: record setup readiness and request timestamps on the first run.

Using "investigate-first". API latency increased after deployment, with no increase in request volume.

Expected outcome:

The strongest evidence points to a new downstream wait. Compare dependency timing before and after deployment to confirm the responsible transition.

Using "investigate-first". A report is empty for one customer, but logs and sample data are unavailable.

Expected outcome:

No cause can be supported yet. The exact blocker is missing customer-specific input, filtering decisions, and report generation output.

Security Audit

Safe
v1 • 8/12/2026 Open versioned report

The two prompt-only files contain no executable code, external access, credential handling, or suspicious instructions. No static or semantic security findings were identified.

2
Files scanned
22
Lines analyzed
0
Review items
0
False positives ignored
No confirmed security findings were detected by the latest completed static and semantic audit. This does not prove the skill has no side effects.
Audited by: codex
Share & cite this report

Share the versioned assessment report, neutral badge, embed card, and citations. Skillstore reports evidence without deciding whether this Skill is safe.

Open versioned report
Security Assessment

Copy report link

https://skillstore.io/skills/juliusbrussee-investigate-first/audits/1?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report

Markdown badge

[![Skillstore security assessment](https://skillstore.io/badges/skills/juliusbrussee-investigate-first/security.svg)](https://skillstore.io/skills/juliusbrussee-investigate-first?utm_source=security_passport_badge)

HTML badge

<a href="https://skillstore.io/skills/juliusbrussee-investigate-first?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/juliusbrussee-investigate-first/security.svg" alt="Skillstore security assessment" loading="lazy"></a>

Embed card

<iframe src="https://skillstore.io/embed/skills/juliusbrussee-investigate-first.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>
Academic citations (APA · BibTeX · CFF)

APA citation

juliusbrussee. (2026). investigate-first security audit report (audit version 1) [Author version unspecified]. Skillstore. https://skillstore.io/skills/juliusbrussee-investigate-first/audits/1

BibTeX citation

@techreport{juliusbrussee-juliusbrussee-investigate-first-2026, author = {juliusbrussee}, title = {investigate-first security audit report (audit version 1)}, institution = {Skillstore}, year = {2026}, number = {1}, url = {https://skillstore.io/skills/juliusbrussee-investigate-first/audits/1}, note = {Author version unspecified} }

CITATION.cff

cff-version: 1.2.0 message: "If you use this Skill, cite its author and this versioned security audit report." title: "investigate-first security audit report (audit version 1)" version: "unspecified" type: report authors: - name: "juliusbrussee" date-released: "2026-08-12" url: "https://skillstore.io/skills/juliusbrussee-investigate-first/audits/1" identifiers: - type: other value: "skillstore:juliusbrussee-investigate-first:audit:1" description: "Skillstore immutable audit report identifier"

Skillstore Score

Why this score Evidence Confidence: Medium
55
Architecture
85
Maintainability
87
Content
65
Community
91
Spec Compliance

What You Can Build

Diagnose an Intermittent Test Failure

Organize observations and rank likely mechanisms before modifying test or product code.

Investigate a Performance Regression

Trace changed inputs and state transitions to identify evidence supporting the slowdown.

Clarify an Ambiguous Defect Report

Separate reported symptoms from assumptions and define the next useful evidence to collect.

Try These Prompts

Review a Failure 
Investigate this failure before proposing edits: [symptom]. Separate observations from assumptions and list the evidence needed next.
Rank Root Cause Hypotheses 
Analyze [failure details]. Rank plausible causes by supporting evidence, conflicting evidence, and the cheapest test that could disprove each cause.
Trace a Regression 
Investigate the regression from [baseline] to [current behavior]. Trace inputs, state transitions, ownership boundaries, and outputs without changing code.
Build an Evidence-Backed Diagnosis 
Evaluate [system context and evidence]. Identify the most credible mechanism, prove each causal step, and name the exact blocker if evidence remains insufficient.

Best Practices

  • Provide exact symptoms, timestamps, reproduction steps, and relevant recent changes.
  • Label observations, assumptions, and hypotheses separately.
  • Prefer inexpensive tests that can quickly disprove the leading explanation.

Avoid

  • Do not treat temporal correlation as proof of causation.
  • Do not edit multiple components before identifying a supported mechanism.
  • Do not continue exploring after evidence proves the cause or identifies the exact blocker.

Frequently Asked Questions

When should I use this skill?
Use it for unknown causes, intermittent failures, performance regressions, or investigations with competing explanations.
Will this skill edit my code?
It will not edit code unless your task explicitly authorizes implementation after diagnosis.
What information should I provide?
Provide symptoms, expected behavior, reproduction steps, logs, timestamps, recent changes, and relevant system boundaries.
How does it rank hypotheses?
It compares each hypothesis with supporting evidence and prioritizes inexpensive tests that could disprove it.
What happens when evidence is insufficient?
The skill identifies the exact missing evidence or access that blocks a credible diagnosis.
Can it investigate performance problems?
Yes. It can trace inputs, state transitions, boundaries, and output timing to identify a supported regression mechanism.

Developer Details

License

MIT

Skillstore revision

r1

Version notice

The author did not declare a version.

Repository

https://github.com/juliusbrussee/caveman/tree/a0109974ea3258a14aadaef1ed1f8ff2837d30d5/skills/investigate-first

Ref

d8fdd465806692db394c1b3f87ab3a021fe733d2

Maintenance freshness

8/13/2026

Usage

0 downloads · 0 views

File structure

📁 agents/

📄 openai.yaml

📄 SKILL.md

More from juliusbrussee

View all

Related Skills

View all