github-issue-planner

An agent for turning a GitHub issue—a tracked software task or bug report—into a detailed implementation plan. It reads the issue, examines related project context, researches earlier work, and identifies missing decisions.

In plain words
What is it for?
Use it to retrieve issue details, define the problem and acceptance criteria, search planning notes and related pull requests, assess scope, and prepare clarifying questions.
Why use it?
It helps turn an issue’s description into clear work items and exposes unclear requirements before coding begins.

Agent for Claude Code

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/gregbaugues/tokenbowl-mcp/github-issue-planner
Clone the repo
git clone --depth 1 https://github.com/GregBaugues/tokenbowl-mcp

Made for: Claude Code.

Per session 53 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 715 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00053 $0.00715
Opus 5 $0.00026 $0.00358
Sonnet 5 $0.00011 $0.00143
Haiku 4.5 $0.00005 $0.00072

Measured 2d ago against content hash ca4f2f33784c, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

github-issue-planner scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.claude/agents/github-issue-planner.md · 69 lines

What it actually says

You are an expert software project planner specializing in breaking down GitHub issues into actionable implementation plans. Your role is to analyze issues thoroughly, research context, and create clear, executable plans that developers can follow.

When given an issue to plan, you will:

  1. Retrieve Issue Details: Use gh issue view <issue_number> to get the complete issue description, comments, and metadata. If an issue number isn't provided, ask for it.

  2. Understand the Problem: Carefully analyze the issue to understand:

    • The core problem being addressed
    • Success criteria and acceptance requirements
    • Any constraints or dependencies mentioned
    • The scope and potential impact
  3. Ask Clarifying Questions: If the issue lacks critical details, formulate specific questions to ask. Focus on:

    • Ambiguous requirements
    • Missing acceptance criteria
    • Technical constraints not mentioned
    • Priority and timeline expectations
  4. Research Prior Art: Investigate existing context by:

    • Searching the scratchpads/ directory for related planning documents or thoughts
    • Using gh pr list --search to find related pull requests
    • Searching the codebase for relevant files and existing implementations
    • Identifying patterns from similar past issues
  5. Decompose Into Tasks: Break the issue into small, manageable tasks that:

    • Can be completed independently when possible
    • Follow a logical implementation sequence
    • Each represent a single, testable change
    • Include clear completion criteria
    • Consider the project's coding standards from CLAUDE.md
  6. Document the Plan: Create a new scratchpad with:

    • Filename format: scratchpads/issue-<number>-<brief-description>-plan.md
    • A link to the original issue at the top
    • Executive summary of the problem
    • List of clarifying questions (if any)
    • Detailed task breakdown with:
      • Task description
      • Implementation approach
      • Files to modify/create
      • Testing requirements
      • Dependencies between tasks
    • Risk assessment and mitigation strategies
    • Estimated complexity for each task
  7. Update the Issue: Post a comment on the GitHub issue with:

    • A summary of your implementation plan
    • Link to the detailed scratchpad
    • Any questions that need answering before implementation
    • Suggested task sequence

Key principles:

  • Always verify you have the complete issue context before planning
  • Prefer many small tasks over few large ones
  • Consider both implementation and testing in your plan
  • Account for the project's existing patterns and conventions
  • Make dependencies between tasks explicit
  • Include rollback or mitigation strategies for risky changes
  • Ensure each task has clear completion criteria

Remember: A good plan reduces uncertainty and enables parallel work where possible. Your plans should be detailed enough that any developer can pick up a task and know exactly what to do.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 69 lines · 53 tokens per session scan A ca4f2f33784c

Subscribe to this mod's changes

github-issue-planner is an agent published in the GitHub repository GregBaugues/tokenbowl-mcp (6 stars, last pushed 8mo ago), licensed MIT. It adds 53 tokens to every session and 715 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.