claude

agentic-dev

Module 16 (Agentic Dev) — the loop for building product with agents: an orchestrator that acts as product owner, separate contracts for dev, QA, security and release, a deploy gate that decides what ships without a human, a source-weighted live backlog, two-tier verification and simulated-user audits. Ships a stack blueprint with a keyless, self-configuring deploy. An incident in one project becomes a rule in all of them.

3 on GitHubby Ocean One Venturesv0.9.6PolyForm-Noncommercial-1.0.0updated 18 days agoproduct-tech-delivery
/

Plugin agentic-dev from marketplace founder-os in steffenmaas/founder-os. Commands read from the repo's manifest — 18 days ago.

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 (12)

  • dev-checkin

    Produces the Agentic Dev check-in — what was built, how it was verified, what is open, blockers, roadmap drift, quality signals and decisions needed from the human. Use this skill when the user says "check-in", "status", "what happened today", "weekly review" or any natural variant. Module 16 (Agent

  • dev-dashboard

    Publishes and refreshes the live development dashboard — a shareable artifact showing product version and scope progress, roadmap, backlog top items, the bundle in flight, changes waiting at the deploy gate with preview URLs, the decision queue, and the last deploys with health state. Use this skill

  • dev-learn

    Captures a learning in docs/learnings, and pushes generalisable ones upstream to the Founder OS repository as a pull request. Use this skill when the user says "write that down", "we learned something", "capture this learning", "send learnings upstream", or after anything surprising, slow, or broken

  • dev-loop

    Runs one unit of work through the Agentic Dev loop under the Dev Agent contract — orient, spec, plan, build, verify, hand to review, ship, learn. Use this skill when the user says "build this feature", "implement", "work on", "fix this bug", "start the loop" or any natural variant. Module 16 (Agenti

  • dev-metrics

    Measures development speed, failure rate and code quality from git history — commit cadence, batch size, change failure rate proxy, rework rate, test ratio, hotspots and infrastructure maturity. Use this skill when the user says "how fast are we shipping", "dev metrics", "code quality", "DORA" or an

  • dev-onboard

    Sets up Agentic Dev in a repository — writes PRODUCT.md, ROADMAP.md, CLAUDE.md, AGENTS.md, spec/ADR/learning/release templates, the single test-then-deploy workflow (CI/security templates stay opt-in), and the managed .founder-os directory. Use this skill when the user says "set up agentic dev", "on

  • dev-product

    Maintains the intent hierarchy — writes and updates PRODUCT.md and ROADMAP.md, checks that every roadmap item traces to the product, and prepares product version cuts. Use this skill when the user says "update the roadmap", "what should we build next", "cut a version", "is this on the roadmap" or an

  • dev-review

    Reviews a diff with fresh context against the acceptance criteria under the QA Agent contract — correctness, edge cases, security, test integrity, suppressed errors. Reports findings, never fixes them. Use this skill when the user says "review this", "check my changes", "is this ready to merge" or a

  • dev-security

    Runs the security gate over a diff or a whole repository under the Security Agent contract — secrets, injection, authorisation, CI permissions, prompt-injection surface, dependencies, data protection. Use this skill when the user says "security check", "is this safe", "audit the repo", "security rev

  • dev-ship

    Takes an approved change to production under the Release Agent contract — precondition check, full verification chain, migration reverse test, PR with rollback plan, preview check, merge, deploy, health watch. Use this skill when the user says "ship it", "deploy", "release", "take it live" or any na

  • dev-spec

    Writes or reworks a specification for one unit of work — problem, goal, non-goal, traceability to the product, affected files, executable acceptance criteria, verification step, risks and rollback. Use this skill when the user says "write a spec", "spec this out", "what exactly should we build" or a

  • dev-stack

    Sets a project up on a proven stack blueprint — the building blocks, the keyless deploy script, the deploy workflow and the deployment documentation, filled in with this project's real values. Use this skill when the user says "set up deployment", "keyless deploy", "WIF", "wire up Firebase", "which

Subagents (6)

  • builder

    Implements ONE increment of an approved spec or plan step — writes the code and its tests, runs the touched scope's checks, commits named files. Use for every build dispatch in the loop; one increment per dispatch, never a whole feature. Writes code but never reviews, approves, merges, or deploys it

  • orchestrator

    Acting product owner between the founder's decisions — grooms the backlog, forms bundles, dispatches every piece of implementation to other subagents, takes the small decisions itself, and reports in four lines. Use to run the development loop. Writes backlog, specs, check-ins and the dashboard; nev

  • planner

    Turns an approved spec into a stepwise implementation plan without writing code. Each step independently committable, each with its own check. Use between spec approval and implementation.

  • reviewer

    Reviews a diff with fresh context against acceptance criteria. Use proactively after any non-trivial code change, before opening a PR. Sees only the diff and the criteria, never the author's reasoning.

  • security-auditor

    Security review of a diff or repository — secrets, injection, authorisation, CI permissions, dependencies, prompt-injection surface, data protection. Use before releases and on any change to auth, data access, CI, or dependencies.

  • verifier

    Runs a project's verification chain and reports the results with evidence — lint, typecheck, tests, build, plus change-specific checks. Use before any unit of work is reported as done.

Hooks (2)

  • PreToolUse

    ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/guard-bash.sh

  • PostToolUse

    ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/scan-secrets.sh

View source on GitHub

Maintain agentic-dev?

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 pluginslisted