DshamirDshamir/slm-forge/skills/forge-preflight/SKILL.md

forge-preflight

Skill1 starsChanged 6 months agoMITView on GitHub

Single-shot environment readiness check. Runs FIRST in every forge invocation, before plan generation. Hard-fails with a structured "what to fix" report if any required credential, tool, quota, or service is missing. The whole forge is gated on this passing — there's no point planning a forge if the environment can't execute it.

Reads credentials

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

Distinctive:
All hard-failing — no recoverable.
line 71

≈ “Common” is an estimate: at least 1 in 5 of the 308,232 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

forge-preflight

When this fires

Phase position: PREFLIGHT (-2) — before ANALYZE and PLAN. The first thing /slm-forge does. If it fails, the operator sees a fix list and the forge never advances.

What it checks

#	Check	Pass criterion
1	Host tools	docker, jq, curl, git, python3 all on PATH
2	AWS forge creds	FORGE_AWS_ACCESS_KEY_ID + secret resolve from vault or env; sts get-caller-identity returns user `intellident-forge-provisioner`
3	S3 bucket reachable	`aws s3 ls s3://<YOUR_S3_BUCKET>/forge/` succeeds
4	HF token	resolves from vault or env; whoami returns namespace; has `repo:create` scope
5	Anthropic API key	resolves from `aiProviders` collection; `messages.create` with 5-token max succeeds
6	G+VT vCPU quota	sufficient headroom for the planned instance type; surfaces stop-instance-id if dental-prod is consuming the quota
7	AZ availability	g5.xlarge offered in at least one AZ in ca-central-1
8	Disk headroom	≥ 5 GB free in /tmp (for staging) and ≥ 20 GB in working dir (for HF cache)

Output

On PASS:

{
  "status": "pass",
  "next_phase": "ANALYZE",
  "resolved_creds": {
    "aws_user": "intellident-forge-provisioner",
    "hf_namespace": "Nexless",
    "anthropic_key_id": "sk-ant-api03-Av...",  // truncated, never full key
    "s3_bucket": "<YOUR_S3_BUCKET>",
    "s3_prefix": "forge/"
  },
  "available_quota": {
    "g_vt_vcpu_total": 8,
    "g_vt_vcpu_in_use": 0,
    "g_vt_vcpu_available": 8,
    "blocking_instances": []  // or [{"id":"i-04b...", "name":"dental-prod", "type":"g4dn.2xlarge", "vcpus":8}]
  }
}

On FAIL:

{
  "status": "fail",
  "blockers": [
    {"check":"hf_token_scope","detail":"token has no repo:create scope","fix":"regenerate token at https://huggingface.co/settings/tokens with write+create scope"},
    {"check":"g_vt_vcpu_quota","detail":"0 vCPUs available; dental-prod (g4dn.2xlarge=8 vCPU) is consuming the quota","fix":"`aws ec2 stop-instances --instance-ids i-04b483e8b944738d6` OR request quota increase"}
  ]
}

Inputs

(none — environmental check; doesn't read manifest)
Optional env: FORGE_REGION (default ca-central-1)

Outputs

Stdout: status JSON
Side effect: caches resolved creds to /tmp/forge-creds.env (mode 600) for reuse by downstream skills

Failure modes

All hard-failing — no recoverable. The forge cannot proceed without:

AWS creds (provision will fail)
HF token (register will fail)
Anthropic key (synth + plan-fit will fail)
vCPU quota (provision will fail)

Usage

bash slm-forge/skills/forge-preflight/run.sh
# Prints JSON to stdout. Exit 0 on pass, 1 on fail.

Download now (ZIP)

In this repository29 other files it gives its agents.
- slm-forgeSkill
- forge-analyzeSkill
- forge-architectSkill
- forge-auditSkill
- forge-bootstrapSkill
- forge-card-validatorSkill
- forge-curateSkill
- forge-estimateSkill
- forge-evalSkill
- forge-ingest-dbSkill
- forge-ingestSkill
- forge-intakeSkill
- forge-monitorSkill
- forge-plan-fitSkill
- forge-planSkill
- forge-prepSkill
- forge-provisionSkill
- forge-publishSkill
- forge-quantizeSkill
- forge-registerSkill
- forge-reportSkill
- forge-resumeSkill
- forge-shapeSkill
- forge-smoketestSkill
- forge-sourceSkill
- forge-statusSkill
- forge-synthSkill
- forge-teardownSkill
- forge-trainSkill

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