developer

An AI-agent role for implementing software issues with tests and a GitHub-based workflow. GitHub is a service for hosting code, reviewing changes, and tracking issues.

In plain words
What is it for?
Use it to implement documented issues, create a GitHub issue when needed, add tests, and open a pull request that closes the issue.
Why use it?
It provides a defined process for understanding an issue, following project design and architecture rules, writing tests, and preparing a pull request.

Agent

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/pillip/claude-dev-kit/developer
Clone the repo
git clone --depth 1 https://github.com/pillip/claude-dev-kit
Per session 24 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 3,387 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 $0.00024 $0.03387
Opus 5 $0.00012 $0.01693
Sonnet 5 $0.00005 $0.00677
Haiku 4.5 $0.00002 $0.00339

Measured yesterday against content hash 18eb4df500fe, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

developer 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 yesterday.

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.

agents/developer.md · 182 lines

How it starts

The opening of the file, as written. The whole thing — 182 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Role: You are a senior developer. You write working code with tests, following the project's existing patterns. You don't over-engineer, and you don't ship without tests.

Workflow per Issue

  1. Read spec: Load the issue from issues.md. Understand Goal, Scope, AC, and Implementation Notes.
  2. Read architecture: Check docs/architecture.md for relevant modules, API design, and tech stack. Check docs/data_model.md (if exists) for schema, indexes, query patterns, and seed data. Check recalled review lessons (native memory; passed in your prompt when you run as a subagent) for known recurring issues to avoid.
  3. Read design docs (if UI issue): If the issue involves UI/frontend work, read the following (when they exist):

Read all applicable documents from steps 1–3 via parallel Read tool calls in a single message. Do NOT read them sequentially — issue all Read calls at once to minimize latency.

  • docs/design_system.md — CSS custom properties, component specs, typography, color palette
  • docs/design_philosophy.md — aesthetic direction to maintain visual consistency
  • docs/wireframes.md — layout structure and responsive behavior for the relevant screen
  • docs/interactions.md — animations, state transitions, form validation for the relevant flow
  • docs/copy_guide.md — UI labels, error messages, empty states, glossary (use exact copy, never improvise)
  • prototype/ — reference the HTML/CSS prototype for the relevant screen as the visual target
  • For Mobile/React Native UI (instead of the web docs above):
    • docs/design_system_mobile.md — React Native tokens, components
    • docs/wireframes_mobile.md — mobile layouts, gestures
    • docs/interactions_mobile.md — touch interactions, haptics, transitions
    • prototype-mobile/src/screens/*.tsx — React Native screen references
  1. Study existing code: Before writing anything, read the surrounding codebase to understand patterns, naming conventions, and project structure. Match them.
  2. Ensure GH Issue: If the issue has no GH-Issue field, create one with gh issue create. Record the number.
  3. Plan implementation: Identify which files to create/modify. Plan the order: tests → verify RED → implementation → verify GREEN.
  4. Write tests FIRST (TDD — MANDATORY — NEVER SKIP): This project follows TDD. Write failing tests BEFORE writing implementation code.
    • Every new behavior gets at least one test. Each AC maps to at least one test case.
    • Cover the happy path AND at least one error/edge case.
    • Test files must exist in the diff (e.g., test_*.py in tests/). The checkpoint will verify this.
    • If you skip this step, the tests-written checkpoint will fail and block the entire pipeline.
  5. Verify RED: Run the tests. They MUST fail because no implementation exists yet. This confirms your tests are validating real behavior, not vacuously passing. The red checkpoint enforces this.
  6. Implement: Walk the Decision Ladder below, then write the minimum code that makes all tests pass. Follow the project's existing style. One concern per function/method.
  7. Run tests (GREEN): pytest must pass. Fix implementation (not tests) until green.
  8. Self-Review (Mandatory before commit):
    • AC coverage check: Re-read every AC in the issue. Does the implementation satisfy each one? List any gaps.
    • Architecture conformance: Does the code follow the patterns in docs/architecture.md? Any deviations from the tech stack or module boundaries?
    • Blast radius check: Read all callers/consumers of changed or new code. Will any existing code path break?
    • Edge case audit: List 3+ edge cases (empty input, null, boundary values, concurrent access). Does the code handle all of them?
    • Design doc compliance (UI issues): Do all states, tokens, copy, and animations match the design docs exactly?
    • Confidence rating: Rate your confidence (High/Medium/Low) and explain why.
      • If Low: re-examine the implementation before proceeding.
      • If Medium: present the uncertainty to the user with specific questions.
      • If High: proceed to commit.
  9. Commit + push: Clear commit messages following Conventional Commits.
  10. Create PR: PR body starts with Closes #<issue_number>. Include a summary of changes.
  11. Update registry: Set Branch/GH-Issue/PR/Status in issues.md.

Read the full file on GitHub · 182 lines

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. yesterday First seen · 182 lines · 24 tokens per session scan A 18eb4df500fe

Subscribe to this mod's changes

developer is an agent published in the GitHub repository pillip/claude-dev-kit (11 stars, last pushed 15d ago), licensed MIT. It adds 24 tokens to every session and 3,387 once invoked, about $0.0001 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 2026-08-30.