Agent skill
debug
Guide systematic debugging through hypothesis generation and verification
Install this agent skill to your Project
npx add-skill https://github.com/thoreinstein/opencode-config/tree/main/skill/debug
SKILL.md
Structured Debugging Workflow
Current Time: !date
System: !uname -a
Guide systematic debugging through hypothesis generation, investigation, and verification. Documents findings to Obsidian for knowledge retention.
Input
- Problem statement: symptoms, error messages, reproduction steps
- Context: environment, recent changes, frequency
- Optional: initial hypotheses or suspected areas
Investigation Strategy
Launch parallel investigation tracks:
Track 1: Codebase Exploration (explore agent)
- Trace code paths related to error messages/symptoms
- Find error handling and logging in affected areas
- Identify recent changes to suspect code
- Map data flow through affected components
Track 2: Code Analysis (inferred agent: go/frontend/postgres)
- Deep analysis of suspect code paths
- Identify potential failure modes
- Review error handling completeness
- Check for race conditions, edge cases, nil handling
Track 3: External Research (librarian agent)
- Search for known issues matching error signatures
- Find similar bugs in dependency issue trackers
- Research common causes for the symptom pattern
Debugging Framework
1. Problem Definition
- What is the expected behavior?
- What is the actual behavior?
- What changed recently?
- Is it reproducible? How?
2. Hypothesis Generation
Generate ranked hypotheses based on:
- Probability (how likely is this the cause?)
- Testability (how easy to verify/falsify?)
- Recent changes (correlation with deployments)
- Error signatures (what do the errors suggest?)
3. Investigation Plan
For each hypothesis:
- What evidence would confirm it?
- What evidence would refute it?
- What's the fastest way to test?
4. Evidence Collection
- Log analysis
- Metric correlation
- Code inspection
- Reproduction attempts
- Binary search (git bisect)
5. Root Cause Identification
- Distinguish symptoms from causes
- Identify contributing factors
- Verify fix addresses root cause
Output
Write to Obsidian via obsidian_append_content at:
$OBSIDIAN_PATH/Debugging/YYYY-MM-DD-HHMM-issue-title.md
Note:
$OBSIDIAN_PATHmust be a vault-relative path (e.g.,Projects/myapp), set per-project via direnv. Theobsidian_append_contenttool expects paths relative to the vault root.
Document Structure
Use this template for the Obsidian document:
@~/.config/opencode/templates/debug-session.md
Behavior
- Parse problem statement to extract symptoms and context
- Infer appropriate code agent from error context
- Launch explore, librarian, and inferred code agent in parallel
- Generate ranked hypotheses based on findings
- Create investigation plan for top hypotheses
- Document evidence and investigation log
- Write debug session to Obsidian via
obsidian_append_contentwith auto-generated filename:YYYY-MM-DD-HHMM-issue-title.md
$ARGUMENTS
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
seo
Implement SEO optimizations for specified pages or entire site
runbook
Generate operational runbooks for services, procedures, or incident response
file-organizer
Intelligently organizes your files and folders across your computer by understanding context, finding duplicates, suggesting better structures, and automating cleanup tasks. Reduces cognitive load and keeps your digital workspace tidy without manual effort.
brainstorming
You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.
ultrathink
Ultra-deep multi-perspective analysis for complex architectural and strategic decisions
commit
Analyze unstaged changes and suggest atomic commit groups with messages
Didn't find tool you were looking for?