developer-pack plugin
Core developer workflows: specs, TDD, debugging, merge conflicts, architecture, PR review, commits, unit and e2e tests, secrets, Docker, SQL, migrations, API SDKs, durable agent workflows, CI security, agent guardrails
How to install the developer-pack plugin
Run both commands inside Claude Code. The first adds the marketplace once; the second installs the plugin.
/plugin marketplace add milad-ahmd/agent-skills
/plugin install developer-pack@milad-agent-skillsWhat's inside developer-pack
21 parts read from the plugin's own files.
Skills 21
Claude loads it when the task matches- agent-guardrails
Installs and tests guardrails that stop a coding agent from running destructive or leaking shell commands. Adds a Claude Code PreToolUse hook on Bash that denies or asks before force pushes to main/master, pushes and commits on main, git reset --hard and git clean -fdx, rm -rf of /, ~, .. or paths outside the project, reading .env files, SSH keys and cloud credentials, curl or wget piped to bash, DROP TABLE and TRUNCATE through psql/mysql/sqlite3, chmod -R 777, and --no-verify. Parses chains, pipes, sudo/env wrappers, bash -c and $(...) and ships a self-test table. Merges into settings.json with a backup and a clean uninstall, and gives the same hook for Cursor and Codex or a plain rule list for other agents. Use when the user asks to add git guardrails or safety hooks to Claude Code, block dangerous or destructive commands, stop the agent from force pushing, deleting files with rm -rf or reading .env secrets, protect the main branch from the agent, or set up a PreToolUse hook that blocks commands.
- agent-memory-keeper
Carries a coding agent's context across sessions with two files. MEMORY.md holds durable project lessons (verified commands, conventions, decisions, gotchas, user preferences); it is loaded at the start of work and updated whenever the user corrects the agent. HANDOFF.md holds the state of the current task (goal, status checklist, exact next step, failed attempts, open questions, branch and uncommitted files); it is written when a session ends mid-task and checked against git when the next session resumes. Use when the user says "remember this", "don't forget", "forget that note" or "what do you remember", says "write a handoff", "I'm running out of context", "continue this in a new session", "pick up where we left off", or "hand this to another agent or a teammate", and at the start of any session in a repo with MEMORY.md or HANDOFF.md.
- api-to-sdk
Generates a clean, typed, idiomatic client SDK (TypeScript, Python, Go, or Dart) from an OpenAPI/Swagger spec, Postman collection, or existing backend routes, with auth, errors, retries, pagination, docs, and tests. Use when the user asks to generate an API client, SDK, or typed fetch layer, or to wrap a REST API.
- architecture-doctor
Finds the architecture problems that cost a team the most time and proposes small, incremental refactors instead of a rewrite. A bundled script measures coupling (fan-in, fan-out, instability), circular dependencies, god files, hubs and layering violations, then weights each finding by git churn, so code that changes every week ranks above code nobody touches. It also finds hidden coupling, files that keep changing in the same commits without importing each other. The agent reads the worst offenders and returns 3-5 ranked refactors, each with evidence, a target shape, a first safe step and a way to verify nothing broke. Use when the user asks to improve the architecture, decide what to refactor first, untangle spaghetti or circular dependencies, split a god class or a file that is too big, fix tight coupling or wrong-way layer imports, pay down tech debt, or draw better module or domain boundaries.
- bug-detective
Debugs code bugs systematically. Builds a minimal repro that fails on demand, logs each hypothesis and how it was tested so the investigation doesn't go in circles, narrows the cause by binary search over code, input or history, and pins down the change that introduced a regression with a git bisect runner that checks both ends first, handles timeouts and flaky repros, and always restores HEAD. Then writes a failing test, fixes the root cause instead of the symptom, and looks for the same bug elsewhere. Use when the user asks to debug their code, fix a bug, find out why a test or function fails, throws an error or returns the wrong result, track down a regression ("it used to work", "which commit broke this"), run git bisect, or chase a flaky or intermittent failure that only happens sometimes.
- changelog-from-git
Generates polished, human-readable release notes and CHANGELOG entries from a git commit range, grouping changes by impact and crediting contributors. Use when the user asks for release notes, a changelog, "what changed since vX", or to draft a GitHub release.
- codebase-explainer
Produces an onboarding guide for an unfamiliar repository from its real import graph. A bundled script parses Python, JS/TS and Go imports to find entry points, hub files that many others depend on, import cycles, orphans and external packages, then emits a Mermaid architecture diagram and a self-contained interactive HTML dependency graph. The guide covers the architecture overview, how a request or data flows from the API to the database, key modules and the files to handle with care, each backed by numbers. Use when the user asks to explain a codebase, onboard to a new or just-cloned repo, map the codebase or its modules, draw an architecture or dependency diagram, see what imports what, or asks "where do I start?".
- commit-craft
Turns messy working-tree changes into clean, atomic Conventional Commits, splitting unrelated changes into separate commits with clear messages. Use when the user asks to commit, write a commit message, split changes into commits, or clean up git history before a PR.
- dependency-doctor
Audits project dependencies for vulnerabilities, outdated and unused packages, and produces a safe, ordered upgrade plan with breaking-change notes. Use when the user asks to update dependencies, fix npm/pip audit warnings, check for vulnerable packages, or plan a major-version upgrade.
- dockerize-anything
Generates a production-ready, secure, small Dockerfile, .dockerignore and docker-compose.yml for any project by detecting its stack, then builds and verifies the image. Use when the user asks to dockerize, containerize, add Docker or compose, or reduce Docker image size.
- durable-agent-workflows
Makes long-running, multi-step Python AI agent jobs crash-proof and resumable with the open-source rewindable library, so a failure at step 37 resumes there instead of re-running and re-paying for steps 1 to 36. Wraps each model call, tool call and side effect in a journaled step (ctx.step, ctx.llm), moves time, randomness and network reads behind ctx so replays match, puts emails, payments and deploys behind human approval, caps spend with a Budget, and proves it by killing a run and resuming it. Includes an AST audit script that flags calls breaking durable replay. Use when the user builds an LLM research pipeline, a batch job over many documents or an agent whose tools send email, pay or deploy, and wants it to survive crashes, resume after an exception without paying again, pause hours or days for a human approval, rewind a run to a step after fixing a bug, or asks for durable execution, checkpointing or human-in-the-loop for a Python agent.
- e2e-flow-tester
Finds a web app's critical user flows (sign up, log in, checkout, password reset, creating and editing the core object, search, settings) and writes Playwright end-to-end tests that drive a real browser through them, using the project's existing Playwright or Cypress setup when there is one, then de-flakes the tests and adds a CI job. Use when the user asks to add e2e or end-to-end tests, set up Playwright, test a signup, login or checkout flow in the browser, write smoke tests for user journeys, or catch UI regressions before deploys.
- env-secrets-guard
Scans a repository and its git history for leaked secrets (API keys, tokens, passwords, private keys), fixes .gitignore and env handling, and guides safe key rotation. Use before pushing, before open-sourcing a repo, or when the user asks to check for leaked keys or secrets.
- gh-actions-auditor
Audits GitHub Actions workflows and composite actions for CI/CD supply-chain and injection risks (actions not pinned to a commit SHA, missing or write-all GITHUB_TOKEN permissions, pull_request_target "pwn request" checkouts, script injection from issue and PR titles, secrets inherit and curl-to-shell installs), then pins SHAs and hardens the YAML with the user's approval. Use when the user asks to secure, harden or audit GitHub Actions or .github/workflows, pin actions to SHAs, or check a CI workflow for injection.
- merge-conflict-resolver
Resolves git merge, rebase, cherry-pick, revert and stash pop conflicts by working out what each side meant and keeping both intents, instead of picking ours or theirs. A read-only script reports the operation in progress and what "ours" and "theirs" mean for it (they swap during a rebase), each conflicted file and its kind (content, add/add, modify/delete, rename, binary), every hunk's ours, base and theirs text, and the commits on each side that touched the file. Lockfiles such as package-lock.json, yarn.lock or Cargo.lock are regenerated with the package manager, never hand-merged, and a check mode confirms no conflict markers are left before git add and continue. Use when the user asks to resolve or fix merge conflicts, says a rebase or cherry-pick stopped with CONFLICT, sees <<<<<<< HEAD markers in a file, asks whether to keep ours or theirs, has a conflicted package-lock.json, or wants help finishing a stuck merge or rebase.
- migration-wizard
Plans and executes large codebase migrations incrementally and safely, such as JavaScript→TypeScript, Jest→Vitest, Next.js Pages→App Router, Vue 2→3, React class→hooks, CommonJS→ESM, or any framework or major-version upgrade, verifying each step with build and tests. Use when the user asks to migrate, convert, upgrade, or port a codebase or module.
- plan-griller
Grills the user about a planned feature or code change before any code is written. It reads the relevant code and docs first so every question is specific to the repo, asks one decision at a time with a recommended answer, works through scope, behaviour, edge cases, data, failure modes and rollout, then writes a SPEC.md with numbered requirements, Given/When/Then acceptance criteria tied to them and a test plan, checked by a linter that flags vague words and untraced requirements. Use when the user asks to grill or interview them about a feature, pressure-test an implementation plan, write a spec or PRD for a change, or turn a rough feature request into testable requirements before coding.
- pr-reviewer-pro
Reviews a pull request or local branch diff like a senior engineer, covering correctness, security, performance, naming, and missing tests, and produces prioritized, actionable comments (optionally posted to GitHub). Use when the user asks to review a PR, a diff, a branch, or "check my changes before I push".
- sql-optimizer
Explains slow SQL queries in plain English, reads EXPLAIN plans, and suggests concrete rewrites and indexes for PostgreSQL, MySQL, and SQLite, including ORM-generated N+1 problems. Use when the user shares a slow query, an EXPLAIN output, asks why a query is slow, or wants index recommendations.
- tdd-loop
Runs strict test-driven development one behaviour at a time. Picks the next smallest behaviour from the spec or task list, writes one failing test, makes the smallest change that passes it, then refactors with the tests green. A guard script records each step and refuses the usual shortcuts. It rejects a red step that fails for the wrong reason (syntax, import or compile error, missing function, no tests ran) and code written before any failing test. It also rejects a test that was edited, deleted, skipped or focused with .only to reach green. Works with pytest, unittest, Jest, Vitest, go test, cargo test, RSpec, Maven and Gradle. Use when the user asks to do TDD, work test-first, write the test first, follow red green refactor, build a feature test-driven from a spec or user story, or stop the agent from changing tests to make them pass.
- test-gap-hunter
Finds the riskiest untested code in a project and writes meaningful tests for it using the project's existing test framework and conventions. Use when the user asks to improve coverage, add tests, find untested code, or harden a module before a refactor.
Using developer-pack in Claude Code
Skills need no command. Claude reads them when your request matches what they describe.
Update, turn off or remove
- Update the plugin (shell)
- claude plugin update developer-pack@milad-agent-skills
- Refresh the marketplace
- /plugin marketplace update milad-agent-skills
- Turn it off for now
- /plugin disable developer-pack@milad-agent-skills
- Remove it
- /plugin uninstall developer-pack@milad-agent-skills
Skills in developer-pack
- agent-guardrailsmilad-ahmd/agent-skills
Installs and tests guardrails that stop a coding agent from running destructive or leaking shell commands. Adds a Claude Code PreToolUse hook on Bash that denies or asks before force pushes to main/master, pushes and commits on main, git reset --hard and git clean -fdx, rm -rf of /, ~, .. or paths outside the project, reading .env files, SSH keys and cloud credentials, curl or wget piped to bash, DROP TABLE and TRUNCATE through psql/mysql/sqlite3, chmod -R 777, and --no-verify. Parses chains, pipes, sudo/env wrappers, bash -c and $(...) and ships a self-test table. Merges into settings.json with a backup and a clean uninstall, and gives the same hook for Cursor and Codex or a plain rule list for other agents. Use when the user asks to add git guardrails or safety hooks to Claude Code, block dangerous or destructive commands, stop the agent from force pushing, deleting files with rm -rf or reading .env secrets, protect the main branch from the agent, or set up a PreToolUse hook that blocks commands.
4 files5stars - agent-memory-keepermilad-ahmd/agent-skills
Carries a coding agent's context across sessions with two files. MEMORY.md holds durable project lessons (verified commands, conventions, decisions, gotchas, user preferences); it is loaded at the start of work and updated whenever the user corrects the agent. HANDOFF.md holds the state of the current task (goal, status checklist, exact next step, failed attempts, open questions, branch and uncommitted files); it is written when a session ends mid-task and checked against git when the next session resumes. Use when the user says "remember this", "don't forget", "forget that note" or "what do you remember", says "write a handoff", "I'm running out of context", "continue this in a new session", "pick up where we left off", or "hand this to another agent or a teammate", and at the start of any session in a repo with MEMORY.md or HANDOFF.md.
4 files5stars - architecture-doctormilad-ahmd/agent-skills
Finds the architecture problems that cost a team the most time and proposes small, incremental refactors instead of a rewrite. A bundled script measures coupling (fan-in, fan-out, instability), circular dependencies, god files, hubs and layering violations, then weights each finding by git churn, so code that changes every week ranks above code nobody touches. It also finds hidden coupling, files that keep changing in the same commits without importing each other. The agent reads the worst offenders and returns 3-5 ranked refactors, each with evidence, a target shape, a first safe step and a way to verify nothing broke. Use when the user asks to improve the architecture, decide what to refactor first, untangle spaghetti or circular dependencies, split a god class or a file that is too big, fix tight coupling or wrong-way layer imports, pay down tech debt, or draw better module or domain boundaries.
5 files5stars - bug-detectivemilad-ahmd/agent-skills
Debugs code bugs systematically. Builds a minimal repro that fails on demand, logs each hypothesis and how it was tested so the investigation doesn't go in circles, narrows the cause by binary search over code, input or history, and pins down the change that introduced a regression with a git bisect runner that checks both ends first, handles timeouts and flaky repros, and always restores HEAD. Then writes a failing test, fixes the root cause instead of the symptom, and looks for the same bug elsewhere. Use when the user asks to debug their code, fix a bug, find out why a test or function fails, throws an error or returns the wrong result, track down a regression ("it used to work", "which commit broke this"), run git bisect, or chase a flaky or intermittent failure that only happens sometimes.
3 files5stars - codebase-explainermilad-ahmd/agent-skills
Produces an onboarding guide for an unfamiliar repository from its real import graph. A bundled script parses Python, JS/TS and Go imports to find entry points, hub files that many others depend on, import cycles, orphans and external packages, then emits a Mermaid architecture diagram and a self-contained interactive HTML dependency graph. The guide covers the architecture overview, how a request or data flows from the API to the database, key modules and the files to handle with care, each backed by numbers. Use when the user asks to explain a codebase, onboard to a new or just-cloned repo, map the codebase or its modules, draw an architecture or dependency diagram, see what imports what, or asks "where do I start?".
3 files5stars - durable-agent-workflowsmilad-ahmd/agent-skills
Makes long-running, multi-step Python AI agent jobs crash-proof and resumable with the open-source rewindable library, so a failure at step 37 resumes there instead of re-running and re-paying for steps 1 to 36. Wraps each model call, tool call and side effect in a journaled step (ctx.step, ctx.llm), moves time, randomness and network reads behind ctx so replays match, puts emails, payments and deploys behind human approval, caps spend with a Budget, and proves it by killing a run and resuming it. Includes an AST audit script that flags calls breaking durable replay. Use when the user builds an LLM research pipeline, a batch job over many documents or an agent whose tools send email, pay or deploy, and wants it to survive crashes, resume after an exception without paying again, pause hours or days for a human approval, rewind a run to a step after fixing a bug, or asks for durable execution, checkpointing or human-in-the-loop for a Python agent.
4 files5stars - e2e-flow-testermilad-ahmd/agent-skills
Finds a web app's critical user flows (sign up, log in, checkout, password reset, creating and editing the core object, search, settings) and writes Playwright end-to-end tests that drive a real browser through them, using the project's existing Playwright or Cypress setup when there is one, then de-flakes the tests and adds a CI job. Use when the user asks to add e2e or end-to-end tests, set up Playwright, test a signup, login or checkout flow in the browser, write smoke tests for user journeys, or catch UI regressions before deploys.
4 files5stars - env-secrets-guardmilad-ahmd/agent-skills
Scans a repository and its git history for leaked secrets (API keys, tokens, passwords, private keys), fixes .gitignore and env handling, and guides safe key rotation. Use before pushing, before open-sourcing a repo, or when the user asks to check for leaked keys or secrets.
2 files5stars - gh-actions-auditormilad-ahmd/agent-skills
Audits GitHub Actions workflows and composite actions for CI/CD supply-chain and injection risks (actions not pinned to a commit SHA, missing or write-all GITHUB_TOKEN permissions, pull_request_target "pwn request" checkouts, script injection from issue and PR titles, secrets inherit and curl-to-shell installs), then pins SHAs and hardens the YAML with the user's approval. Use when the user asks to secure, harden or audit GitHub Actions or .github/workflows, pin actions to SHAs, or check a CI workflow for injection.
3 files5stars - merge-conflict-resolvermilad-ahmd/agent-skills
Resolves git merge, rebase, cherry-pick, revert and stash pop conflicts by working out what each side meant and keeping both intents, instead of picking ours or theirs. A read-only script reports the operation in progress and what "ours" and "theirs" mean for it (they swap during a rebase), each conflicted file and its kind (content, add/add, modify/delete, rename, binary), every hunk's ours, base and theirs text, and the commits on each side that touched the file. Lockfiles such as package-lock.json, yarn.lock or Cargo.lock are regenerated with the package manager, never hand-merged, and a check mode confirms no conflict markers are left before git add and continue. Use when the user asks to resolve or fix merge conflicts, says a rebase or cherry-pick stopped with CONFLICT, sees <<<<<<< HEAD markers in a file, asks whether to keep ours or theirs, has a conflicted package-lock.json, or wants help finishing a stuck merge or rebase.
3 files5stars - plan-grillermilad-ahmd/agent-skills
Grills the user about a planned feature or code change before any code is written. It reads the relevant code and docs first so every question is specific to the repo, asks one decision at a time with a recommended answer, works through scope, behaviour, edge cases, data, failure modes and rollout, then writes a SPEC.md with numbered requirements, Given/When/Then acceptance criteria tied to them and a test plan, checked by a linter that flags vague words and untraced requirements. Use when the user asks to grill or interview them about a feature, pressure-test an implementation plan, write a spec or PRD for a change, or turn a rough feature request into testable requirements before coding.
4 files5stars - pr-reviewer-promilad-ahmd/agent-skills
Reviews a pull request or local branch diff like a senior engineer, covering correctness, security, performance, naming, and missing tests, and produces prioritized, actionable comments (optionally posted to GitHub). Use when the user asks to review a PR, a diff, a branch, or "check my changes before I push".
2 files5stars - tdd-loopmilad-ahmd/agent-skills
Runs strict test-driven development one behaviour at a time. Picks the next smallest behaviour from the spec or task list, writes one failing test, makes the smallest change that passes it, then refactors with the tests green. A guard script records each step and refuses the usual shortcuts. It rejects a red step that fails for the wrong reason (syntax, import or compile error, missing function, no tests ran) and code written before any failing test. It also rejects a test that was edited, deleted, skipped or focused with .only to reach green. Works with pytest, unittest, Jest, Vitest, go test, cargo test, RSpec, Maven and Gradle. Use when the user asks to do TDD, work test-first, write the test first, follow red green refactor, build a feature test-driven from a spec or user story, or stop the agent from changing tests to make them pass.
3 files5stars - api-to-sdkmilad-ahmd/agent-skills
Generates a clean, typed, idiomatic client SDK (TypeScript, Python, Go, or Dart) from an OpenAPI/Swagger spec, Postman collection, or existing backend routes, with auth, errors, retries, pagination, docs, and tests. Use when the user asks to generate an API client, SDK, or typed fetch layer, or to wrap a REST API.
5stars - changelog-from-gitmilad-ahmd/agent-skills
Generates polished, human-readable release notes and CHANGELOG entries from a git commit range, grouping changes by impact and crediting contributors. Use when the user asks for release notes, a changelog, "what changed since vX", or to draft a GitHub release.
5stars - dependency-doctormilad-ahmd/agent-skills
Audits project dependencies for vulnerabilities, outdated and unused packages, and produces a safe, ordered upgrade plan with breaking-change notes. Use when the user asks to update dependencies, fix npm/pip audit warnings, check for vulnerable packages, or plan a major-version upgrade.
5stars - migration-wizardmilad-ahmd/agent-skills
Plans and executes large codebase migrations incrementally and safely, such as JavaScript→TypeScript, Jest→Vitest, Next.js Pages→App Router, Vue 2→3, React class→hooks, CommonJS→ESM, or any framework or major-version upgrade, verifying each step with build and tests. Use when the user asks to migrate, convert, upgrade, or port a codebase or module.
5stars - sql-optimizermilad-ahmd/agent-skills
Explains slow SQL queries in plain English, reads EXPLAIN plans, and suggests concrete rewrites and indexes for PostgreSQL, MySQL, and SQLite, including ORM-generated N+1 problems. Use when the user shares a slow query, an EXPLAIN output, asks why a query is slow, or wants index recommendations.
5stars - test-gap-huntermilad-ahmd/agent-skills
Finds the riskiest untested code in a project and writes meaningful tests for it using the project's existing test framework and conventions. Use when the user asks to improve coverage, add tests, find untested code, or harden a module before a refactor.
5stars - commit-craftmilad-ahmd/agent-skills
Turns messy working-tree changes into clean, atomic Conventional Commits, splitting unrelated changes into separate commits with clear messages. Use when the user asks to commit, write a commit message, split changes into commits, or clean up git history before a PR.
5stars - dockerize-anythingmilad-ahmd/agent-skills
Generates a production-ready, secure, small Dockerfile, .dockerignore and docker-compose.yml for any project by detecting its stack, then builds and verifies the image. Use when the user asks to dockerize, containerize, add Docker or compose, or reduce Docker image size.
5stars
developer-pack plugin questions
How do I install the developer-pack plugin?
In Claude Code run /plugin marketplace add milad-ahmd/agent-skills, then /plugin install developer-pack@milad-agent-skills.
What does the developer-pack plugin include?
developer-pack bundles 21 skills.
How do I update or remove the developer-pack plugin?
Update it with claude plugin update developer-pack@milad-agent-skills in your shell, or from the Installed tab of /plugin. Remove it with /plugin uninstall developer-pack@milad-agent-skills.