Skip to content

unxed/f4 / .agents/skills/aif

Set up agent context for a project. Analyzes tech stack, installs relevant skills from skills.sh, generates custom skills, and configures MCP servers. Use when starting new project, setting up AI context, or asking "set up project", "configure AI", "what skills do I need".

FileKindSize
SKILL.mdskill md32.9 KB
references/config-template.yamlconfig9.5 KB
references/update-config.mjsscript16.9 KB

references/config-template.yaml

View on GitHub
# AI Factory Configuration
# This file configures AI Factory behavior for your project.
# All sections are optional — defaults are used when not specified.

# =============================================================================
# Language Settings
# =============================================================================
# Controls the language used in AI-generated content and communication.

language:
  # Language for AI-agent communication (prompts, questions, explanations)
  # Options: en, ru, de, fr, es, zh, ja, ko, pt, it
  # Default: en
  ui: en

  # Language for generated artifacts (plans, specs, documentation)
  # Options: en, ru, de, fr, es, zh, ja, ko, pt, it
  # Default: en (same as ui if not specified)
  artifacts: en

  # How to handle technical terms in translations
  # Options:
  #   - keep: preserve original English terms (API, database, etc.)
  #   - translate: translate where common translation exists
  #   - mixed: translate prose terms but keep code/infrastructure ecosystem terms
  # Default: keep
  technical_terms: keep

# =============================================================================
# Path Configuration
# =============================================================================
# Custom paths for AI Factory artifacts.
# All paths are relative to project root.

paths:
  # Project description file
  # Default: .ai-factory/DESCRIPTION.md
  description: .ai-factory/DESCRIPTION.md

  # Architecture guidelines file
  # Default: .ai-factory/ARCHITECTURE.md
  architecture: .ai-factory/ARCHITECTURE.md

  # Detailed documentation directory used by /aif-docs
  # README.md remains the landing page in the project root.
  # Change this if you want docs pages under another folder like
  # documentation/, handbook/, or site/docs/.
  # Default: docs/
  docs: docs/

  # Roadmap file
  # Default: .ai-factory/ROADMAP.md
  roadmap: .ai-factory/ROADMAP.md

  # Research notes file.
  # Explicit /aif-explore ultra derives named bundles at
  # <parent directory>/research/<english-topic-slug>/; no second path key is needed.
  # Default: .ai-factory/RESEARCH.md
  research: .ai-factory/RESEARCH.md

  # Top-level project rules file (axioms)
  # Default: .ai-factory/RULES.md
  rules_file: .ai-factory/RULES.md

  # Fast-mode plan file
  # Used by /aif-plan fast for a single quick plan.
  # Keep this separate from paths.plans unless you want fast plans
  # to live next to full named plans.
  # Default: .ai-factory/PLAN.md
  plan: .ai-factory/PLAN.md

  # Named plan artifacts directory
  # /aif-plan full stores named plan files here.
  # /aif-plan ultra stores named directories with index.md + phase files here.
  # In git-aware mode this is typically branch-scoped.
  # In no-git mode it stores slug-based plan files.
  # Default: .ai-factory/plans/
  plans: .ai-factory/plans/

  # Fix plan file
  # Default: .ai-factory/FIX_PLAN.md
  fix_plan: .ai-factory/FIX_PLAN.md

  # Security ignore-state file
  # Default: .ai-factory/SECURITY.md
  security: .ai-factory/SECURITY.md

  # Knowledge references directory
  # Default: .ai-factory/references/
  references: .ai-factory/references/

  # Self-improvement patches directory
  # Default: .ai-factory/patches/
  patches: .ai-factory/patches/

  # Evolution logs directory
  # Default: .ai-factory/evolutions/
  evolutions: .ai-factory/evolutions/

  # Reflex loop state directory
  # Default: .ai-factory/evolution/
  evolution: .ai-factory/evolution/

  # Specs directory (archived plans)
  # Default: .ai-factory/specs/
  specs: .ai-factory/specs/

  # Rules directory
  # Default: .ai-factory/rules/
  rules: .ai-factory/rules/

  # QA artifacts root directory
  # /aif-qa stores change-summary, test-plan, and test-cases here.
  # /aif-qa-check stores qa-check.md here and binds execution results to
  # the tested revision/build identifier plus deterministic test-cases
  # digests.
  # /aif-qa-check agent also stores root-level agent-context.md and
  # agent-history.md with non-sensitive cross-QA setup facts and
  # recurring learnings; per-run details stay in branch qa-check.md.
  # Both commands use the same deterministic, filesystem-safe derived
  # branch slug as a subdirectory: <qa>/<branch-slug>/
  # The slug uses mandatory 40-character safe-slug truncation plus a short
  # hash suffix for collision resistance (see skills/aif-qa/SKILL.md for
  # the exact algorithm).
  # Default: .ai-factory/qa/
  qa: .ai-factory/qa/

  # Archive directory for completed plans and roadmap snapshots.
  # /aif-archive moves finished plans here (retaining original file/directory
  # names, including any sequential NNNN_ prefix). Subdirectories:
  #   archive/plans/   — archived plan files and ultra bundle directories
  #   archive/roadmap/ — dated roadmap snapshots
  # Referenced by /aif-archive (owner), /aif-plan, /aif-implement,
  # /aif-verify, /aif-improve.
  # Default: .ai-factory/archive/
  archive: .ai-factory/archive/

# =============================================================================
# Workflow Settings
# =============================================================================
# Controls AI Factory workflow behavior.

workflow:
  # Automatically create .ai-factory/ directories when missing
  # Default: true
  auto_create_dirs: true

  # Plan ID format for new plans
  # Options: slug, sequential   (timestamp and uuid are reserved — see below)
  # Default: slug (derived from branch or task description)
  #
  # Only `slug` and `sequential` are active right now. `timestamp` and `uuid`
  # are accepted by `/aif-plan` but currently behave like `slug` (an INFO log
  # is emitted). They are reserved for a future filename format that the
  # branch-based consumers (`/aif-implement`, `/aif-improve`, `/aif-verify`,
  # `/aif-rules-check`) will know how to discover.
  #
  # `sequential` writes full plans as `<NNNN>_<stem>.md` and ultra bundles as
  # `<NNNN>_<stem>/index.md` under `paths.plans/`, where:
  #   - `stem` is the canonical filename stem chosen by `/aif-plan`:
  #       1. `HANDOFF_BRANCH_NAME` with `/` replaced by `-`  (Handoff mode)
  #       2. the generated git branch name with `/` replaced by `-`
  #       3. the description slug (when no branch is created)
  #     Branch-based consumers reproduce the same stem at lookup time.
  #   - NNNN = max(existing 4-digit prefix) + 1, zero-padded to 4 digits;
  #     an empty plans directory starts from `0001`.
  # Numbering is derived from existing numbered plan files and numbered
  # directories whose index.md contains the exact ultra discovery marker.
  # Other numbered directories do not consume IDs. If the highest-numbered
  # plan artifact is deleted or moved elsewhere,
  # that number can be reused on the next run; keep plans in place if you
  # rely on stable cross-references. The format is bounded: 0001..9999.
  # `/aif-plan` errors out instead of writing `10000_…` so consumer globs
  # stay aligned. Branch names stay un-numbered (the prefix lives on the full
  # filename or ultra directory). Force-disabled when Handoff owns the branch
  # (HANDOFF_BRANCH_PREPARED=1) — Handoff requires the plan identifier to equal
  # the branch name. Fast plans (`paths.plan`) and fix plans
  # (`paths.fix_plan`) are single files and ignore this setting.
  plan_id_format: slug

  # Whether the setup/analyze flow updates ARCHITECTURE.md
  # Default: true
  analyze_updates_architecture: true

  # Whether /aif-architecture updates ROADMAP.md
  # Default: true
  architecture_updates_roadmap: true

  # Default verification mode
  # Options: strict, normal, lenient
  # Default: normal
  verify_mode: normal

# =============================================================================
# Git Settings
# =============================================================================

git:
  # Whether this project uses git-aware workflows
  # If false:
  #   - /aif-plan full and /aif-plan ultra do not create branches
  #   - /aif-review and /aif-verify do not assume a base branch exists
  #   - merge/push guidance is skipped
  # Default: auto-detected from the repository (.git presence)
  enabled: true

  # Default branch for diff/review/merge targets
  # Examples: main, master, develop, trunk, 2.x
  # Default: auto-detected from git metadata, fallback main
  base_branch: main

  # Automatically create feature branches for plans
  # Applies only when git.enabled = true
  # If false, /aif-plan full and /aif-plan ultra still create plan artifacts
  # but stay on the current branch (using slug-named artifacts in no-git mode)
  # Default: true when git is enabled
  create_branches: true

  # Branch name prefix for new features
  # Applies only when create_branches = true
  # Default: feature/
  branch_prefix: feature/

  # Skip push prompt after /aif-commit creates a commit
  # If true:
  #   - /aif-commit ends after successful local commit
  #   - no "Push to remote?" prompt is shown
  # Default: false
  skip_push_after_commit: false

# =============================================================================
# Rules Configuration
# =============================================================================
# Configure the rules hierarchy for this project.

rules:
  # Base rules file (project-specific conventions)
  # This file is created by /aif during project setup
  # Default: .ai-factory/rules/base.md
  base: .ai-factory/rules/base.md

  # Area-specific rules files
  # These are created by /aif-rules when you add area-specific conventions
  # Uncomment and add paths as needed:
  # api: .ai-factory/rules/api.md
  # frontend: .ai-factory/rules/frontend.md
  # backend: .ai-factory/rules/backend.md
  # database: .ai-factory/rules/database.md
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall