Diagnose code failures by ranking falsifiable hypotheses and fixing root causes.
npx skills add https://github.com/shiwenwen/hope-agent --skill ha-debug
Please help me install this Agent Skill. Skill: ha-debug Source: https://github.com/shiwenwen/hope-agent/tree/main/skills/ha-debug Command: npx skills add https://github.com/shiwenwen/hope-agent --skill ha-debug
Debugging often stalls when developers jump to the first plausible explanation instead of working from evidence. This Skill provides a disciplined decision process for code failures, regressions, crashes, flaky behavior, and bad output: characterize the failure, bound the fault, rank falsifiable hypotheses, fix the smallest root cause, and prove the failing path.
Ask the agent to debug an intermittent test failure by finding and fixing the root cause from the failing logs.
High-intent search queries and answers about installing and using this skill.
Characterize whether the failure is timing-sensitive, data-specific, or platform-specific, then bound the fault across async ordering, locks, and concurrency boundaries. Rank one or two falsifiable hypotheses and run the cheapest discriminating check before editing code.
Trace the smallest credible path through inputs, state transitions, persistence, and outputs, comparing what crosses each component boundary. Patch only the smallest ownership boundary that restores the contract, and stop after two failed attempts to re-challenge your assumptions.
Characterize the failure from logs, fixtures, persisted state, and code paths instead of live reproduction, and state the evidence gap explicitly. Never mark an unreproduced hypothesis as confirmed; report the strongest substitute evidence and remaining uncertainty.
No. Passing compilation alone does not prove a runtime bug is fixed. Use a focused regression test, a deterministic failing command, a before/after database or log query, or manual reproduction that would have failed before the fix.
Pause only when progress requires inaccessible user state, an external system change, a destructive action, or a product decision. Do not invent data or broaden scope into speculative rewrites before the fault is proven.
ilove323Diagnose software failures and generate a structured debugging report.
Comfy-OrgFixes bugs as two commits so CI proves the regression test catches the bug.
VedaAstroImplement pytest, Vitest, and Playwright testing strategies for Veda projects.
MovchanetsDiagnose software bugs and performance issues through hypothesis-driven experimentation.
