investigate

Investigate bugs and code areas with evidence-graded findings recorded in a structured case file.

Updated Jul 26, 2026
One-click install
npx skills add https://github.com/chrissstellee/Movix --skill investigate-chrissstellee
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: investigate
Source: https://github.com/chrissstellee/Movix/tree/main/.agents/skills/investigate
Command: npx skills add https://github.com/chrissstellee/Movix --skill investigate-chrissstellee

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Debugging and incident analysis often rely on guesswork, lost context, and unverified assumptions. This Skill enforces a disciplined forensic workflow that reconstructs what happened from available evidence, grades every finding as Confirmed, Deduced, or Hypothesized, and produces a persistent case file another engineer can pick up cold. ## Core Features & Use Cases - Evidence-graded investigation: Anchors analysis in a Confirmed stronghold, traces causality backward from symptoms, and runs refutation passes before confirming hypotheses. - Persistent case file: Initializes a structured Markdown case file (evidence inventory, timeline, hypotheses, backlog, conclusion with confidence level) that survives interruptions and supports follow-up sessions. - Dual-mode operation: Handles both symptom-driven defect chasing (tickets, logs, stack traces) and symptom-free code area exploration to build a mental model before making changes. - Use Case: A user reports a failing escrow release on a Stellar-based trade platform. Activate the skill with the error message; it anchors on the confirmed error, maps the evidence perimeter across logs and version control, traces the root cause through the Soroban contract call chain, and delivers a case file with a High-confidence conclusion and fix direction. ## Quick Start Ask the agent to investigate a bug by providing a ticket ID, error message, log file path, or code area name, for example: investigate why the escrow release transaction fails after delivery confirmation.

Frequently Asked Questions about investigate

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I investigate a bug with a structured root cause analysis workflow?▼

Provide a ticket ID, error message, log path, or problem description. The skill anchors on one confirmed piece of evidence, maps the evidence perimeter, forms and tests hypotheses with refutation passes, and stops at a confidence-rated diagnosis with fix direction.

How to trace the root cause of an incident from logs and stack traces?▼

The skill records the log path and time window, then traces causality backward from the symptom to producing conditions. It cross-references logs, version control, and system events to reconstruct a timeline, citing every finding with path:line or commit hash.

Can I use this to understand unfamiliar code before making changes?▼

Yes. Exploration mode works without a symptom: it maps inputs, outputs, and dependencies, scans frequent terms, and filters control flow to build a mental model of the code area, ending with a verification plan instead of a root cause.

Does the investigation skill modify or fix the code it analyzes?▼

No. Investigation stops at the diagnosis; implementation is out of scope. For trivial fixes like an off-by-one or missing null check, it may include a one-line suggestion or draft diff in the report, then routes follow-up work to other workflows.

What happens when there is not enough evidence to find the root cause?▼

The skill switches to evidence-light mode: it marks the case accordingly, builds a prioritized data-collection backlog, and records exactly what evidence is needed to proceed. Missing evidence is documented as a finding with its impact and how to obtain it.