agentleFS
Sign inStart free
DshamirDshamir/slm-forge/skills/forge-intake/SKILL.md

forge-intake

Skill1 starsChanged 6 months agoMITView on GitHub

Collects model requirements (goal, domain, corpus_ref, target_use, license preference, language, constraints) interactively or from an --auto-spec JSON file. Validates against D-005 (300M parameter ceiling for the v1 size class) before letting the run proceed. Entry point for the legacy v1 pipeline; v2 runs use forge-analyze + forge-plan instead.

We split this skill into the rules it gives an agent

3 rules. 0 are common, 0 are distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 4 edited versions we read keep these.

  • 0

    Distinctive

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    If manifest.spec is already populated, exit 0 (already intaken)

    line 16
  2. Very specific:

    Write manifest.spec and advance state to ARCHITECT

    line 21
  3. Very specific:

    If manifest.spec is non-empty, exits 0 without re-prompting.

    line 33

“Common” is measured: kept by at least 6 in 10 of the 4 edited versions of this file we could read. “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

forge-intake

When this fires

Phase position: INTAKE — entry point for the legacy v1 manifest-driven flow. Precedes forge-architect. Skipped on v2 runs (forge-analyze populates analysis.json directly from the input).

What it does

  1. If manifest.spec is already populated, exit 0 (already intaken)
  2. With --auto-spec <path>, read the JSON and validate required keys
  3. Without auto-spec, ask the operator interactively for: goal, domain, corpus_ref, target_use, license_preference, language, constraints
  4. Hard-validate constraints.max_params <= 300_000_000 (D-005 ceiling)
  5. Write manifest.spec and advance state to ARCHITECT

Inputs

  • $1 = forge-id (required)
  • --auto-spec <path> = JSON file with the spec fields (optional)
  • stdin (interactive mode when --auto-spec absent)

Outputs

  • manifest.spec.{goal, domain, corpus_ref, target_use, license_preference, language, constraints}
  • manifest.state.current_phase = ARCHITECT

Idempotency

If manifest.spec is non-empty, exits 0 without re-prompting.

Failure modes

ExitReason
0spec populated (success or already-done)
1missing required spec field OR max_params > 300_000_000 (D-005 violation)
64no forge-id provided

External resources

  • stdin (interactive) or filesystem (--auto-spec JSON)
  • Local manifest only — no AWS / HF / Anthropic calls

Cost class

free — pure local I/O.

Depends on

None — entry point.

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.

Reviews can't be read right now.

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.

Check your setup, free