fmind

agent-mcp

Configure MCP servers for Antigravity, Codex, OpenCode, Claude, and Copilot using stdio or remote transport at workspace or user scope, with source review and safe secret handling.

agent-project

Bootstrap shared project instructions and skills for Antigravity, Codex, OpenCode, Claude, and Copilot, while keeping each tool's MCP, settings, and custom-agent files in native locations.

Author, validate, publish, or install Agent Skill packages. Use for SKILL.md frontmatter, metadata, scripts, references, discovery paths, CLI checks, and reviewed installs; route third-party trust to skill-security-review.

chezmoi

Canonical chezmoi dotfiles setup — the source-attribute naming grammar, Go templates, age-encrypted secrets, and the edit-source then apply/diff workflow. Use when adding, editing, or debugging a managed dotfile.

containerize

Build minimal, non-root OCI images — `ko` for Go or a distroless multi-stage Dockerfile — then scan, sign, and SBOM them. Use when containerizing or packaging an app for deployment.

d2

Create, theme, validate, and export D2 diagrams when an existing D2 source or a bespoke standalone composition is a better fit than the default Mermaid workflow.

dependabot

Automated dependency updates using GitHub Dependabot — configuration, ecosystem mapping, and validation.

fmind-visuals

Apply the Fmind visual identity and route slide or diagram work to Slidev, Mermaid, LikeC4, or D2. Use for Fmind talks, decks, article diagrams, architecture visuals, and www.fmind.dev assets.

git-add-commit-push

Stage, commit (Conventional Commits), and push in one flow, auto-healing lefthook pre-commit and pre-push failures. Use when committing and pushing work end-to-end.

github-actions

Canonical GitHub Actions CI/CD that runs the same repository `mise run all` format, check, test, and build gate as local hooks, plus CD deploy templates. Use when setting up or editing repository workflows.

github-pull-request

Create or update a GitHub pull request with a structured What / Why / How / Test-plan body. Use when opening or updating a PR for the current branch.

github-repository

Configure a GitHub repository's metadata (description, homepage, topics) and solo-developer settings via `gh`, derived from the codebase. Use when setting up or tidying a repo's GitHub settings.

go-stack

Canonical Go development stack for structuring and building projects with standard tooling, scaffolding, packages, CLIs/TUIs, web (GOTH), and ADK agents. Use for any Go project, library, or application.

k8s-local

Create and manage local Kubernetes clusters (k3d or kind) and deploy to them with kubectl, helm, helmfile, and skaffold. Use for local k8s cluster setup, dev loops, and debugging.

lefthook

Canonical lefthook git-hooks setup — pre-commit (format, check, secret scan) and pre-push (test), each delegating to `mise run` tasks. Use when configuring or debugging git hooks.

mermaid

Create, theme, validate, and embed Mermaid diagrams as the default text-based diagram format for GitHub, Slidev, documentation, and technical communication.

mise

Canonical mise setup — the per-project task vocabulary (install, format, check, test, build, watch) and pinned toolchain that hooks and CI reuse. Use when defining `mise.toml` tasks or pinning tools.

python-stack

Canonical Python stack with uv, Ruff, ty, pytest, Litestar, Typer, and dependency source. Use for Python projects, libraries, CLIs, agents, packaging, tests, typing, linting, or verifying APIs without importing them.

release

Cut a versioned release — bump semver, generate the changelog with git-cliff, tag, and publish a GitHub release. Use when shipping a new tagged version.

security-scan

Run approved offline Trivy or gitleaks scans on a checkout for vulnerabilities, dependencies, IaC, secret history, licenses, images, and config; triage findings. Use for full-repo scans, not threat modeling.

upgrade-tools

Upgrade every pinned tool and dependency (mise, go.mod, pyproject.toml, GitHub Actions, dprint, ...) to latest stable, one ecosystem at a time with validation. Use when bumping a repo's toolchain or dependencies.

dot-cli

Use the dot CLI (fmind/dotfiles) to verify the environment, pull repos, manage the local k3d cluster, log agent sessions, and prune caches. Use when running or scripting any dot command.

github-triage

Triage, review, and label GitHub pull requests and issues from the terminal with gh-dash, octo.nvim, and gh. Use when reviewing agent-authored PRs, grooming an issue backlog, or deciding which labels to set.

issue-execution

Execute one eligible GitHub issue end to end through cooperative claiming, bounded context, acceptance-driven implementation, repository validation, and an auditable handoff. Use when asked to pick up, implement, finish, or deliver an already-approved issue; reject blocked or human-gated work and never infer authorization to publish, merge, deploy, release, or mutate runtime state.

k8s-review

Review Kubernetes manifests and a specifically verified cluster using bounded, sanitized evidence, then develop evidence-backed optimization hypotheses without changing cluster state. Use for Kubernetes audits, workload health reviews, capacity and scheduling analysis, upgrade-readiness reviews, or optimization assessments; delegate cluster lifecycle and deployment work to k8s-local.

project-backlog

Turn repository-review evidence into a deduplicated, prioritized, dependency-aware backlog. Use when findings already exist and need ordering into issue drafts; require authorization and a confirmed repo before GitHub mutation.

release-audit

Audit an existing GitHub release without changing it. Use when verifying exact-head CI, local and remote tag identity, public release state, or declared release assets such as checksums, SBOMs, signatures, and attestations; do not use it to publish or repair a release.

repository-review

Audit a whole repository across architecture, source, tests, tooling, security, CI/CD, docs, and releases. Use for cross-cutting audits, readiness, or technical-debt reviews with ranked findings — not one diff or release candidate.

Data & Documents Solid

dprint

Canonical dprint setup — the standard formatter for config and markup files (JSON, Markdown, TOML, YAML). Use when configuring or running formatting for these file types.

Data & Documents Solid

python-script

Write standalone single-file Python scripts using PEP 723 inline metadata and uv. Use when creating a quick CLI script that needs dependencies without a full project.

full-review

Run this dotfiles repo's non-destructive health check — every mise task and dot CLI command still work, pinned tools/dependencies are current, and AGENTS.md/README.md/skills/ match the live repo. Use for a periodic whole-repo review.

agent-evaluation

Evaluate stochastic LLM/RAG/model/retrieval/tool agents in trials. Compare baseline/candidate on development/sealed holdouts with calibrated deterministic/model/trace graders; measure reliability, variance, leakage, safety, and cost.

cloud-run

Deploy container apps to Google Cloud Run — ko or Dockerfile images to Artifact Registry, keyless CI via Workload Identity Federation, Secret Manager wiring. Use when shipping a service, web app, or ADK agent to GCP.

diff-review

Inspect changed code in an exact diff, patch, branch, PR, or candidate for correctness defects. Use for pre-merge or self-review, spec compliance, test gaps, and regression risk — not whole-repo audit, release go/no-go, or plan review.

handover

Write a self-contained follow-up prompt carrying this session's decisions, constraints, and verified findings into a fresh agent session. Use when handing unfinished work to a new context window.

Data & Documents Solid

hugo

Canonical Hugo static-site stack with the Hextra docs theme — Hugo Modules, mise tasks, dprint, lefthook, and GitHub Pages deploy. Use for documentation sites, project docs, and static websites.

implementation-plan

Turn accepted requirements into ordered dependency-aware slices and a repository-grounded implementation plan. Use before editing when work spans systems, has migration or rollout risk, needs architecture choices, or should be split.

plan-execution

Execute an accepted implementation plan in bounded, verified slices. Use to coordinate multiple agents and shared-file ownership, resume planned work, or finish scoped tasks without crossing commit, push, deploy, or publication boundaries.

plan-review

Red-team a product, architecture, implementation, migration, or launch plan before execution. Use for pre-mortems, strategy or scope challenge, and independent testing of assumptions, dependencies, and failure modes — not code.

Web & Frontend Solid

product-design-review

Critique customer interface flows using interaction, visual, responsive, keyboard and screen-reader accessibility, content, and runtime evidence. Use for UX audits, design reviews, UI polish, onboarding, empty states, or redesign.

product-loop

Discovery, specification, launch, and learning for a product in one loop. Build-or-stop calls, MVP wedges, demand tests, customer interviews, PRDs, requirements, user journeys, acceptance, positioning, onboarding, rollout, pricing, pivots.

prompt-design

Design production LLM or agent prompt stacks: instructions, tool contracts, examples, outputs, and runtime context. Use for precedence, conflicts, dynamic or untrusted context; prove behavior with agent-evaluation.

repository-history

Reconstruct why tracked code exists from read-only Git history. Trace files, symbols, or lines through introducing commits or pull requests, rationale, reverts, renames, authorship, co-change clues, and issues.

skill-security-review

Audit third-party Agent Skills/extensions for supply-chain security without running them before install/trust. Inspect scripts, hooks, MCP/plugins, hidden instructions, symlinks, credential/network flows, and provenance/licenses.

sops-secrets

Manage repository secrets with sops + age — encrypted files committed to git, memory-only decryption via exec-env, Flux and OpenTofu integration. Use whenever a project needs secrets in version control or at runtime.

systematic-debugging

Diagnose unknown-cause bugs, test/build or auth failures, flakes, and runtime performance regressions. Investigate, reduce, localize, falsify hypotheses, and explain root cause before implementation.

technical-research

Verify unfamiliar framework APIs and architectures from exact local dependency source and current primary docs. Compare installed packages, specs, and versions before choosing an approach; return a recommendation with proof boundaries.

Code & Development Listed

conventional-commit

Write a Conventional Commits subject for the staged changes and commit them. Use when committing staged work with a typed, scoped message.

dot-release

Prepare, tag, and push a versioned fmind/dotfiles release locally to trigger GitHub Actions CD publication.

feature-branch

Create and switch to a new git branch with conventional `<type>/<slug>` naming. Use when starting a piece of work that needs its own branch off main.

project-license

Select and generate the correct LICENSE — MIT for public fmind/fmind-ai repos, otherwise proprietary — and update project manifests. Use when adding or fixing a project's license.

readme-agents

Sync AGENTS.md (for agents) and README.md (for humans) with the codebase's current tools, layout, and usage. Use when either drifts from the project's actual state.

Code & Development Listed

likec4-dsl

Use when working with `.c4`/`.likec4` files or LikeC4 CLI/config questions where exact DSL/CLI syntax is required, especially for strict command/snippet-first answers, validate/export flags, predicates `*`/`_`/`**`, deployment snippets, dynamic views, or relationship extension matching.

Web & Frontend Listed

slidev

Create and present web-based slidedecks for developers using Slidev with Markdown, Vue components, code highlighting, animations, and interactive features. Use when building technical presentations, conference talks, code walkthroughs, teaching materials, or developer decks.