bug-hunt-swarm
Bug Hunt Swarm deploys four read-only sub-agents in parallel to investigate bugs, regressions, and crashes from different angles—reproduction scope, code paths, recent changes, and proof strategies. The main agent then synthesizes findings into ranked hypotheses with supporting evidence and the fastest path to confirm or disprove each theory, all without making code edits.
Bug Hunt Swarm investigates root causes through parallel read-only analysis, ranking likely culprits and fastest proof paths.
AI-generated summary based on this skill's SKILL.md
Decision gist · record as of 2026-03-29
Bug Hunt Swarm investigates root causes through parallel read-only analysis, ranking likely culprits and fastest proof paths. Bug Hunt Swarm deploys four read-only sub-agents in parallel to investigate bugs, regressions, and crashes from different angles—reproduction scope, code paths, recent changes, and proof strategies. The main agent then synthesizes findings into ranked hypotheses with supporting evidence and the fastest path to confirm or disprove each theory, all without making code edits.
Use it when
- Yes.
- Bug Hunt Swarm investigates flaky failures by running multiple agents in parallel, each probing different aspects: reproduction conditions.
Verify before relying
Read SKILL.md below before installing (2 files). Open directory: indexed for reading, not audited.
Install
Dimillian/Skills/bug-hunt-swarm · repository language: Shell
Open directory. Skills are indexed for reading, not audited. Review a skill's body before installing it.
Frequently asked questions
AI-generated answers based on this skill's SKILL.md and metadata
What is Bug Hunt Swarm and how does it find root cause of crashes?
Bug Hunt Swarm is an MIT-licensed investigation framework that deploys four read-only sub-agents in parallel to diagnose bugs, regressions, and crashes from different angles. Each agent examines reproduction scope, code paths, recent changes, and proof strategies independently. The main agent then synthesizes their findings into ranked hypotheses with supporting evidence, helping you identify the root cause without editing code.
Can Bug Hunt Swarm trace a regression to its source using read-only investigation?
Yes. Bug Hunt Swarm traces regressions to their source through read-only investigation. The framework's parallel agents examine code history, recent changes, and execution paths to pinpoint what broke. Findings are ranked by evidence strength, and you get the fastest proof path to confirm or disprove each hypothesis—all without modifying your codebase.
How does Bug Hunt Swarm investigate flaky or intermittent test failures?
Bug Hunt Swarm investigates flaky failures by running multiple agents in parallel, each probing different aspects: reproduction conditions, code paths, timing dependencies, and proof strategies. This multi-angle approach surfaces intermittent patterns and race conditions that single-threaded debugging often misses, then ranks likely causes by evidence.
Why is my code broken and how can Bug Hunt Swarm help diagnose it?
Bug Hunt Swarm diagnoses unexpected behavior and breakage by deploying four read-only agents that examine your code from reproduction scope, path analysis, change history, and proof strategy angles. The framework ranks likely bug causes by evidence strength and provides the fastest confirmation path for each hypothesis, so you understand what broke and why.
Does Bug Hunt Swarm rank bug causes by evidence and proof path?
Yes. Bug Hunt Swarm ranks likely bug causes by evidence and delivers the fastest proof path for each theory. After parallel agents investigate from multiple angles, the main agent synthesizes findings into prioritized hypotheses with supporting evidence, letting you confirm or disprove the most probable causes first.
Can I use Bug Hunt Swarm for production issue debugging without editing code?
Yes. Bug Hunt Swarm is designed for read-only investigation of production issues, crashes, and regressions. The framework deploys parallel agents to analyze reproduction scope, code paths, recent changes, and proof strategies without making edits. You get ranked hypotheses with evidence and the fastest path to root cause, ideal for production diagnostics.
SKILL.md
Rendered from the published skill. Quoted content, verbatim.
Bug Hunt Swarm
Investigate a bug with four read-only sub-agents in parallel, then have the main agent rank the likely causes and recommend the fastest path to prove or fix the issue. This skill is diagnosis-first: do not edit files or implement fixes as part of this workflow.
Step 1: Build the Bug Packet
Start by collecting the smallest useful investigation packet:
- Symptom
- Expected behavior
- Actual behavior
- Reproduction steps, if known
- Scope of impact
- Relevant evidence, such as logs, stack traces, failing tests, screenshots, recent diffs, or environment details
Prefer this source order:
- Direct user description
- Explicit files, stack traces, logs, tests, or screenshots provided by the user
- Current git changes or recent repo history when the bug appears regression-like
- The
(truncated - see the full file via the links below)
File tree — 2 files
bug-hunt-swarm/SKILL.md
bug-hunt-swarm/agents/openai.yaml
Let your AI agent find skills like this
Example. Real query, live index.
You found this page by searching. An agent finds it by wishing: SkillFed indexes 56,283 agent skills by what they can do, searchable in plain language.
wish › “Find the root cause of a bug, crash, or regression without editing code”
Give your agent the search over MCP, or paste the wish link into any chat. No install? Search from any chat →
Related skills
Review Swarm runs four specialized read-only agents in parallel to examine a git diff or file scope for behavioral regressions, security risks, performance issues, and coverage gaps. The main agent then filters and ranks findings by severity, delivering only high-signal issues with a clear fix path—no file edits or automated fixes applied.
Investigate guides you through structured debugging in five phases, from evidence collection to verified fix. It enforces the Iron Law principle—never fix without a confirmed root cause—using pattern analysis, hypothesis testing with a 3-strike escalation rule, and regression coverage. Output includes a machine-readable debug report with the full evidence chain.
This skill enforces disciplined debugging by requiring root cause investigation before any fix attempt. It walks through four phases: isolating the problem through error analysis and reproduction, comparing working patterns against broken code, forming and testing a single hypothesis, and finally implementing the fix while catching architectural issues that multiple failed attempts reveal.
Debug Investigator replaces ad-hoc troubleshooting with structured investigation: capture symptoms, build a reproducible feedback loop, analyze evidence from stacktraces and logs, generate ranked hypotheses, and design bisection strategies. Use it for non-obvious bugs—intermittent failures, performance regressions, or issues requiring systematic root cause analysis—where native debugging falls short.
This skill enforces a disciplined debugging process: establish a tight feedback loop that reproduces the exact symptom, gather evidence across system boundaries, trace data flow to its source, and form testable hypotheses before proposing any fix. By completing each phase sequentially—root cause investigation, pattern analysis, hypothesis testing, and verification—you eliminate guesswork and prevent quick patches from masking deeper issues.
Parallel Debugging applies the Analysis of Competing Hypotheses methodology to organize multi-agent investigation of complex issues. It structures hypothesis generation across six failure mode categories, establishes evidence collection standards with citation requirements, and provides arbitration protocols to identify root causes while avoiding confirmation bias.