discuss-bugs

Investigate GitHub Discussions bug reports and route confirmed bugs to the team.

2|Updated Apr 5, 2026
One-click install
npx skills add https://github.com/witwave-ai/witwave --skill discuss-bugs-witwave-ai
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: discuss-bugs
Source: https://github.com/witwave-ai/witwave/tree/main/.agents/self/piper/.claude/skills/discuss-bugs
Command: npx skills add https://github.com/witwave-ai/witwave --skill discuss-bugs-witwave-ai

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? User-reported bugs in GitHub Discussions require multi-turn investigation — verifying claims against source code, asking clarifying questions, and routing confirmed bugs to the right engineer — which is tedious to do manually on a recurring cadence. ## Core Features & Use Cases - Automated bug triage: Scans the GitHub Discussions Bugs category via the GraphQL API, applies moderation and engagement guards, and tracks per-thread investigation state across ticks. - Code-level verification: Traces reported behavior through the codebase using grep, git blame, git log, and test inspection to confirm or refute bug claims before replying. - Team routing: Queues confirmed bugs in a shared memory file for a dispatcher agent and escalates recurring user misconceptions to the docs owner. - Use Case: A user posts "ww send returns the wrong session id" in the Bugs category; the skill traces the command handler, confirms the regression against a recent commit, replies to the reporter, and queues the bug for the engineering peer. ## Quick Start Ask the agent to check bugs in the GitHub Discussions Bugs category and investigate any new reports.

Frequently Asked Questions about discuss-bugs

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

FAQPage Schema
How do I automatically triage bug reports in GitHub Discussions?▼

Use the GitHub GraphQL API to list discussions in the Bugs category, then verify each report against the source code with grep, git blame, and test inspection. This skill automates that loop on a 15-minute heartbeat, tracking per-thread state across ticks.

How to verify a user bug report against source code?▼

Trace the reported command or endpoint through the code with grep, read the full function body and call graph, then check git blame and recent commits for regressions. Existing tests reveal intended behavior and help distinguish real bugs from environment differences.

Can an AI agent reply to GitHub Discussions comments?▼

Yes, via the addDiscussionComment mutation in the GitHub GraphQL API, optionally passing a replyToId for threaded responses. Guards prevent self-replies, double-posting, and enforce per-thread cooldowns and daily reply ceilings.

What happens to confirmed bugs after investigation?▼

Confirmed bugs are appended to a routing queue file with repro summary, verified commit SHA, suggested target peer, and severity. A separate dispatcher agent reads the queue and assigns the fix; the investigator never writes code or edits other agents' memory.

How are spam or prompt injection comments in bug threads handled?▼

A moderation pre-screen guard pattern-matches every comment for spam, prompt injection, harassment, and similar categories before any reply logic runs. Matched comments are minimized via the API, logged to a moderation file, and excluded from the investigation.