KS
Killer-Skills

gh-discussions-answerer — how to use gh-discussions-answerer how to use gh-discussions-answerer, gh-discussions-answerer setup guide, automated GitHub discussions answering, gh-discussions-answerer vs other AI agent skills, gh-discussions-answerer install, what is gh-discussions-answerer, gh-discussions-answerer alternative, LLM harnesses for GitHub discussions, gh CLI for automated answering

v1.0.0
GitHub

About this Skill

Perfect for Code Review Agents needing automated GitHub discussion response capabilities via LLM harnesses and gh CLI verification. gh-discussions-answerer is a GitHub discussions answerer skill that automates the process of finding and answering discussions using LLM harnesses and the gh CLI

Features

Delegates work to general subagents for efficient processing
Utilizes the gh CLI for 100% code-verified answers
Auto-posts answers immediately after analysis
Discards only uncertain answers to ensure accuracy
Operates in read-only mode until Phase 3, using only GET/query operations

# Core Topics

anntnzrb anntnzrb
[0]
[0]
Updated: 3/7/2026

Quality Score

Top 5%
39
Excellent
Based on code quality & docs
Installation
SYS Universal Install (Auto-Detect)
Cursor IDE Windsurf IDE VS Code IDE
> npx killer-skills add anntnzrb/agents/gh-discussions-answerer

Agent Capability Analysis

The gh-discussions-answerer MCP Server by anntnzrb is an open-source Categories.community integration for Claude and other AI agents, enabling seamless task automation and capability expansion. Optimized for how to use gh-discussions-answerer, gh-discussions-answerer setup guide, automated GitHub discussions answering.

Ideal Agent Persona

Perfect for Code Review Agents needing automated GitHub discussion response capabilities via LLM harnesses and gh CLI verification.

Core Value

Empowers agents to automate answering GitHub discussions with accurate and efficient responses using LLM harnesses, 100% code verification via gh CLI, and auto-posting capabilities, leveraging protocols like GET/query operations for read-only analysis.

Capabilities Granted for gh-discussions-answerer MCP Server

Automating responses to unanswered GitHub discussions
Generating accurate and efficient answers using LLM harnesses
Verifying code responses with 100% accuracy via gh CLI

! Prerequisites & Limits

  • Requires gh CLI for 100% code verification
  • Read-only access until Phase 3
  • Limited to posting 10 answers with 1.5x candidate selection
Project
SKILL.md
4.4 KB
.cursorrules
1.2 KB
package.json
240 B
Ready
UTF-8

# Tags

[No tags]
SKILL.md
Readonly

GitHub Discussions Answerer

All work delegated to general subagents.

Constraints

  • Target: 10 answers → find 1.5x candidates
  • Post ALL verified (may be lower or higher, that's fine)
  • 100% code-verified via gh CLI
  • Discard only uncertain → post everything that passes
  • Auto-post immediately after analysis
  • Read-only until Phase 3 → Phases 1-2 use only GET/query operations; no mutations

Workflow

Phase 1: Discovery

Search for unanswered discussions via GitHub API:

task(subagent_type="general", description="Search unanswered discussions", prompt="
Run these searches to find unanswered discussions. Calculate dates from today.

SEARCH 1 - Last 30 days, Q&A with no replies:
gh api graphql -f query='{
  search(query: \"is:open comments:0 created:>YYYY-MM-DD category:Q&A NOT author:bot\", type: DISCUSSION, first: 100) {
    nodes { ... on Discussion { title number url bodyText repository { nameWithOwner } category { name } } }
  }
}'

SEARCH 2 - Popular repos (stars>100), any unanswered:
gh api graphql -f query='{
  search(query: \"is:open comments:0 stars:>100 NOT author:bot NOT title:RFC\", type: DISCUSSION, first: 100) {
    nodes { ... on Discussion { title number url bodyText repository { nameWithOwner } category { name } } }
  }
}'

SEARCH 3 - Last 14 days, Help/Support categories:
gh api graphql -f query='{
  search(query: \"is:open comments:0 created:>YYYY-MM-DD NOT author:bot NOT title:proposal\", type: DISCUSSION, first: 100) {
    nodes { ... on Discussion { title number url bodyText repository { nameWithOwner } category { name } } }
  }
}'

Merge results, deduplicate by repo#number.

INCLUDE categories (prioritized):
1. Q&A, Questions, Help, Support, Troubleshooting (highest priority)
2. General (if contains a question mark or asks 'how to')
3. Technical, Development, Usage (if asking for help)

HARD SKIP (not answerable with code research):
- Feature requests ('add X', 'would be nice if', 'please implement')
- Timeline/roadmap questions ('when will X', 'release plan', 'ETA')
- Questions only maintainers can answer (prioritization, future plans)
- Bot-created discussions (mvnpm, dependabot, renovate, github-actions)
- RFCs/proposals/announcements/newsletters
- Release notes or changelogs
- Discussions with only emoji or very short body (<20 chars)
- Already answered discussions (comments > 0)

Return exactly 15 candidates: [repo#number] title - category
")

Phase 2: Parallel Analysis

Launch analyses for ALL candidates in parallel (one subagent per candidate):

task(subagent_type="general", description="Analyze discussion", prompt="
Analyze [repo]#[number]: [title]

READ-ONLY: Only use GET/query operations. Do NOT post, create, or mutate anything.

1. Read discussion: gh api repos/OWNER/REPO/discussions/NUMBER
2. Search code: gh search code 'keyword' repo:OWNER/REPO
3. Read files: gh api repos/OWNER/REPO/contents/PATH --jq '.content' | base64 -d
4. Check README/docs for relevant info

VERIFIED answers must provide ACTIONABLE value:
- Code fixes with file:line references
- Configuration solutions with exact syntax
- Workarounds when expected feature is missing
- Technical explanation of WHY something doesn't work
- Links to relevant documentation in the repo

DISCARD if answer would just restate the question:
- 'Feature X doesn't exist' (user knows, that's why they asked)
- 'No timeline available' (useless)
- 'This is a known limitation' (without workaround)

Return VERIFIED: [1-2 sentence answer with actionable fix] OR DISCARD: [reason]
")

Phase 3: Post All Verified

Post ALL verified answers in parallel (one subagent per answer):

task(subagent_type="general", description="Post discussion answer", prompt="
Post answer for [repo]#[number]:

1. Get discussion ID:
   gh api graphql -f query='{ repository(owner:\"X\", name:\"Y\") { discussion(number:N) { id } } }'

2. Post the answer:
   gh api graphql -f query='mutation { addDiscussionComment(input: { discussionId: \"ID\", body: \"ANSWER\" }) { comment { url } } }'

3. Unsubscribe from notifications
   THREAD_ID=$(gh api /notifications --jq '.[] | select(.subject.url | contains(\"OWNER/REPO/discussions/NUMBER\")) | .id')
   gh api -X DELETE /notifications/threads/$THREAD_ID/subscription

Return posted URL.
")

Answer Format

  • 1-2 sentences max
  • File:line if relevant
  • No fluff, no AI-speak
  • Must give user something they can DO, not just confirm what they know

Related Skills

Looking for an alternative to gh-discussions-answerer or building a Categories.community AI Agent? Explore these related open-source MCP Servers.

View All

widget-generator

Logo of f
f

widget-generator is an open-source AI agent skill for creating widget plugins that are injected into prompt feeds on prompts.chat. It supports two rendering modes: standard prompt widgets using default PromptCard styling and custom render widgets built as full React components.

149.6k
0
Design

chat-sdk

Logo of lobehub
lobehub

chat-sdk is a unified TypeScript SDK for building chat bots across multiple platforms, providing a single interface for deploying bot logic.

73.0k
0
Communication

zustand

Logo of lobehub
lobehub

The ultimate space for work and life — to find, build, and collaborate with agent teammates that grow with you. We are taking agent harness to the next level — enabling multi-agent collaboration, effortless agent team design, and introducing agents as the unit of work interaction.

72.8k
0
Communication

data-fetching

Logo of lobehub
lobehub

The ultimate space for work and life — to find, build, and collaborate with agent teammates that grow with you. We are taking agent harness to the next level — enabling multi-agent collaboration, effortless agent team design, and introducing agents as the unit of work interaction.

72.8k
0
Communication