claude

foreman

Splits a task into subtasks, hands each one to a predefined agent type and picks Haiku, Sonnet, Opus or Fable for it by the kind of work, then runs the subtasks in dependency waves and verifies once. Skills dispatch and new-agent; agents scout, mechanic, researcher, builder, writer, reviewer, verifier, architect, debugger and specialist; seven task templates, a plan checker that catches overlapping file ownership before any agent starts, a run log whose stats suggest route changes, a plan-to-Workflow compiler and a scaffolder for new agents and agent plugins. Needs Python 3.12+. Lives in its own repo.

5 on GitHubby Sagar Guptav0.4.0MITupdated yesterdayworkflow
/

Plugin foreman from marketplace sagar-dev-skills in Sagargupta16/claude-skills. Commands read from the repo's manifest — today.

This plugin runs code on your machine — its hooks fire automatically on events like tool use. The exact commands are listed below; read them before installing.

What it installs

Skills (2)

  • dispatch

    Splits a task into subtasks, hands each one to a foreman agent type (scout, mechanic, researcher, builder, writer, reviewer, verifier, architect, debugger, specialist) and picks Haiku, Sonnet, Opus or Fable for each by the kind of work, then runs them in dependency waves, integrates the results and

  • new-agent

    Creates Claude Code subagents and agent plugins in foreman's format. Writes one agent file with the model, effort and tools its kind of work needs, or a whole plugin (manifest, marketplace entry, agents, an optional skill that launches them, README, changelog, license), lints agent files for fields

Subagents (10)

  • architect

    Read-only planner for changes that need judgement. Reads the code a task touches, weighs two or three approaches with their tradeoffs, picks one, and splits the work into lanes with disjoint file ownership, interfaces between lanes and one verify check per lane. Called by foreman dispatch for design

  • builder

    Implements one specified code change inside the files it owns, following the repo's existing style, and proves it with the one check in its brief. Writes tests when the goal asks for them. Called by foreman dispatch for implement subtasks (the build lanes of the feature, refactor and migration templ

  • debugger

    Root-cause investigator. Reproduces a failure with an exact command, quotes the first real error line, narrows it by layer and by bisecting, tests one hypothesis at a time against the evidence that would kill it, and returns the root cause, the evidence, the fix and the exact files the fix touches.

  • mechanic

    Applies edits and runs commands that the brief spells out exactly. Renames with given old and new names, version bumps, find and replace, moving files, applying one proven recipe to each file in a list, regenerating a file with a given command. Called by foreman dispatch for mechanical subtasks (the

  • researcher

    Read-only research lane. Answers one question from primary sources (official docs, API references, changelogs, package registries, source repositories) and returns cited rows of claim, URL, date read and confidence. Called by foreman dispatch for research subtasks (one lane per question or source fa

  • reviewer

    Read-only reviewer for one diff, PR or set of files, through one lens when the brief names it (correctness, security, performance, API design, tests). Returns findings ranked by severity, each with file:line evidence and a concrete fix, and one verdict. Called by foreman dispatch for review subtasks

  • scout

    Read-only locator. Finds the files, symbols, call sites, config keys and tests a task touches, counts matches, and maps a codebase area into a short list of paths with one line each. Called by foreman dispatch for locate subtasks (the map step of the feature, bugfix, refactor, migration, review and

  • specialist

    Takes one long, tangled unit of work end to end. Investigates first, picks the approach, makes the change inside the files it owns and proves it with the check in its brief plus one check of its own. For work where splitting would lose the thread (each step needs everything the last one learned) or

  • verifier

    Read-only claim checker. Reproduces each claim it is handed (a lane's "done", a count, a review finding, a research row, "tests pass") with its own independent command and returns confirmed, refuted or unreproducible per claim with the command and its decisive output. Called by foreman dispatch for

  • writer

    Writes and updates documentation inside the files it owns (README, docs pages, changelog entries, code comments, release notes, migration guides) and checks every command, path, flag and number it writes against the code. Called by foreman dispatch for document subtasks (the docs template and the do

Hooks (1)

  • SubagentStop

    [ "${CLAUDE_PLUGIN_OPTION_CAPTURE_RUNS:-}" = "true" ] || exit 0; in=$(cat); f="${CLAUDE_PLUGIN_ROOT}/hooks/capture.py"; printf '%s' "$in" | python3 "$f" 2>/dev/null || printf '%s' "$in" | python "$f"

View source on GitHub sagargupta.online

More from Sagargupta16/claude-skills

All 20 plugins →

orrery

claude

Architecture, system design and business flow diagrams from a short YAML spec, drawn with the official AWS, Azure, Google Cloud, Kubernetes, GitHub and IBM icons plus 5,500 tech logos, auto-laid out and rendered to SVG, PNG, editable draw.io, React TSX, HTML and design documents. Skill orrery with a bundled CLI; commands /orrery:draw, /orrery:doc, /orrery:import-image, /orrery:review; agents planner, diagram-author, image-importer, patch-proposer and four reviewers (accuracy, conventions, flow, legibility). Lives in its own repo.

1 skill · 4 commands · 8 agents · hooks

5

Everyday dev commands -- commit, review, test, fix bugs, create PRs, check repo status, and inspect PR health. Agents: code-reviewer, debugger. Hook: auto-format. Commands use dynamic context injection for grounded, first-attempt responses.

1 skill · 7 commands · 2 agents · hooks

5

Open source contribution workflow -- upstream sync, CONTRIBUTING.md compliance, code style matching, PR preparation, CLA/DCO signing, AI disclosure, and PR quality bar. Agent: pr-analyzer. Hook: upstream-sync-check. Commands use dynamic context injection.

1 skill · 2 commands · 1 agent · hooks

5

Advanced git operations -- rebasing, cherry-picking, conflict resolution, bisect, stash management, history exploration, and undo strategies. Agent: git-assistant. Hook: commit-lint.

1 skill · 1 command · 1 agent · hooks

5

svg-forge

claude

Design animated SVGs that render everywhere: self-contained README cards, banners, loaders, icons and diagrams that survive GitHub's image proxy, plus React and Motion scenes for apps. Skills animated-svg, motion-scene, svg-audit; agent svg-designer; a generator for 14 kinds with 15 contrast-checked themes, a validator with WCAG contrast rules, a frame renderer and a minifier. Lives in its own repo.

3 skills · 1 agent

5

Context window management -- compaction timing, context rot prevention, branching strategies, and sub-agent isolation patterns. Agent: context-advisor.

1 skill · 1 command · 1 agent

5

Similar plugins

Matt Pocock's agent skills for real engineering (grilling, spec/ticket flows, TDD, code review, domain modelling and more).

39 skills

mattpocock/skills 323,108

superpowers

claudecodex

Core skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques

15 skills · hooks

obra/superpowers 316,836

claude-mem

claudecodex

Persistent memory for Claude Code: the next session starts where the last one left off.

29 skills · 1 MCP server · hooks

thedotmack/claude-mem 116,618

mem0

claudecodex

Cross-session memory and token savings for coding agents.

6 skills · 1 agent · 1 MCP server · hooks

mem0ai/mem0 68,893

context7

claudecodex

Up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your coding agent's context.

1 MCP server

upstash/context7 64,709

Automates browser interactions for web testing, form filling, screenshots, and data extraction

12 skills

vercel-labs/agent-browser 50,373

Maintain foreman?

This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.

claude & codex plugins20 plugins