agentleFS
gerard-labsgerard-labs/superpowers-api-platform/skills/runner-selection/SKILL.md

runner-selection

Skill2 starsChanged 6 months agoMITView on GitHub

Detect the correct command runner for the project — DDEV (`.ddev/`), Make (Makefile with `console`/`tests`/`ci` targets — covers Smile/agency boilerplates where Symfony lives in a sub-dir and Make wraps Docker), Symfony Docker/FrankenPHP (`compose.yaml` with `frankenphp`/`caddy`), generic Docker Compose, or host. Build the right prefix for `bin/console` / `composer` / `phpunit` / `paratest`. Handles monorepos by walking up from `composer.json` to find the orchestration root (Makefile/compose.yaml). The session-start hook does this automatically; this skill documents the manual detection logic and the fallback when the hook hasn't run.

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Always detect before running.

    line 31

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Runner Selection (Symfony)

Use when

  • The session-start hook output is unavailable (manual claude invocation, fresh repo).
  • You need to verify which runner the hook picked.
  • A command works locally but fails because the host environment differs from Docker.
  • The project uses a Makefile-driven workflow (make console, make tests) and you need to know which target to call.

Default workflow

  1. From composer.json (the active app), walk up to the orchestration root — the first parent dir with a Makefile, Makefile-solution, compose.yaml, compose.yml, docker-compose.yml, docker-compose.yaml, or .ddev/. Stop at .git/.
  2. At that root, detect in this priority order:
    • .ddev/ → DDEV
    • Makefile with console / tests / ci targets → Make (most common in agency boilerplates)
    • compose.yaml mentioning frankenphp / dunglas/symfony-docker / caddy → Symfony Docker
    • Any other compose file → generic Compose
    • Otherwise → host
  3. Build the prefix from the prefix table in reference.md.
  4. For Make projects, prefer make <target> over the raw Docker invocation — the target already wraps the right docker compose run --rm or exec call with the right COMPOSE_FILE and env files.

Guardrails

  • Always detect before running; never assume host.
  • A compose.yaml present but no service running ≠ host. Suggest make up (if Make), ddev start (if DDEV), or docker compose up -d (otherwise).
  • For multi-service compose (php + database + redis), the relevant service is usually php or app. Inspect docker compose config --services.
  • For Make projects with a Makefile-solution file, never add new targets to Makefile (framework file, regenerated from the boilerplate). Add to Makefile-solution instead — see skill gerard:makefile-discipline.

Progressive disclosure

  • This file covers the detection logic.
  • reference.md covers the full prefix table per runner, Make convention, monorepo walking, and edge cases.

Output contract

  • A working bin/console / composer / phpunit invocation.
  • A note in the dev report saying which runner was used (so the reviewer can verify).

References

  • reference.md
  • docs/complexity-tiers.md
  • hooks/session-start.sh — the canonical detection (host mode duplicates this logic for manual cases)
  • gerard:makefile-discipline — sister skill for projects driven by make

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.