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

forge-status

Skill1 starsChanged 6 months agoMITView on GitHub

SLM-Forge status helper. Read-only inspection of a forge's manifest, phase history, artifacts, cost breakdown, and error log. Use when the user asks "what's the state of my forge?", "how much has this forge cost?", "what phase is forge <id> at?", or invokes /slm-forge status <id>. Never mutates state — safe to call at any time on any forge regardless of phase.

We split this skill into the rules it gives an agent

4 rules. 0 are common, 3 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 8 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Does NOT mutate the manifest

    line 107
  2. Distinctive:

    Does NOT advance any phase

    line 108
  3. Distinctive:

    Does NOT trigger any sub-skill

    line 109
  4. Very specific:

    Does NOT spend real AWS money beyond a single describe-instance or describe-instance-information call (for liveness)

    line 110

“Common” is measured: kept by at least 6 in 10 of the 8 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-status

Non-phase helper skill. Renders a human-readable status report for a forge without side effects.

Trigger patterns

  • /slm-forge status [forge-id] (forge-id optional; defaults to current-forge pointer)
  • "what's the status of <id>?"
  • "how much has this forge cost?"
  • "show me the forge state"

What it reports

=== Forge: forge-2026-04-23-dental-xyz123 =====================
  created:        2026-04-23T14:30:00Z  (by daniel@mercury)
  phase:          TRAIN (in-progress, 2h 15m)
  updated:        2026-04-23T16:45:12Z  (2 min ago)

Spec:
  goal:           train a dental patient-ed SLM on tiny corpus
  domain:         dental.patient-education
  target_params:  200M
  budget_cap:     $10.00
  max_wall:       24h

Plan:
  base_model:     Qwen/Qwen2.5-0.5B  (Apache-2.0)
  regime:         prune-to-300m
  framework:      unsloth
  chat_template:  qwen2

Estimate:
  instance:       g5.xlarge  @ $1.212/hr
  gpu_hours:      6.50  (confidence=medium)
  total_est:      $9.14

Phase history:
  INTAKE          ✓  2m
  ARCHITECT       ✓  30s
  ESTIMATE        ✓  10s
  BUDGET_GATE     ✓  (passed by user)
  SOURCE          ✓  9m
  CURATE          ✓  25m
  SHAPE           ✓  15m
  PROVISION       ✓  4m
  BOOTSTRAP       ✓  14m
  TRAIN           …  2h 15m  (in-progress)

Artifacts:
  raw_corpus:      s3://.../data/raw/
  curated_corpus:  s3://.../data/curated/
  shaped_corpus:   s3://.../data/shaped/
  checkpoints:     s3://.../checkpoints/  (last: step-1200, loss 1.842)
  final_weights:   (not yet)
  eval_reports:    (not yet)
  quantized:       (not yet)
  hf_repo:         (not yet)
  hf_space:        (not yet)

Compute:
  instance:        i-0abc123def456  (running, SSM online)
  launched:        2026-04-23T15:25:00Z  (1h 20m ago)
  ebs_volume:      vol-0xyz789  (200 GB gp3)

Training:
  pid:             14532  (alive)
  last_loss:       1.842 @ step 1250
  last_heartbeat:  1m ago
  eta:             ~45 min

Cost:
  to_date:         $4.23
  by_phase:        PROVISION $0.07, BOOTSTRAP $0.38, TRAIN $3.78
  cap:             $10.00
  headroom:        $5.77

Gates:
  budget_gate:     ✓ passed  (2026-04-23T14:36:30Z, by user)
  quality_gate:    (pending)

Errors:        (none)
Notes:         (none)

Procedure

skills/forge-status/run.sh <forge-id> (or no arg → uses current-forge pointer from $FORGE_WORK/current-forge.txt).

  1. Load manifest via manifest_load.
  2. Parse + format the sections above.
  3. For cost reconciliation, call compute_aws_cost_to_date (fast estimate, no Cost Explorer hit — status should be cheap).
  4. For training liveness, probe the PID via SSM IF compute_target.instance_id is live AND training_runtime.pid is set (one lightweight SSM exec).
  5. Print and exit 0.

What it doesn't do

  • Does NOT mutate the manifest
  • Does NOT advance any phase
  • Does NOT trigger any sub-skill
  • Does NOT spend real AWS money beyond a single describe-instance or describe-instance-information call (for liveness)

Return contract

Non-phase skill: prints to stdout, exits 0 on success. Does not emit the {"status":"completed","next_phase":"..."} JSON that phase skills do.

Download now (ZIP)

  • Other versions11 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 11

    1. TT-Wang/forge★ 34same name
    2. ggombee/code-forge★ 18same name
    3. lorenzofamiglini/The-Forge-Protocol-Agent★ 15same name
    4. antonailog/forge-skills★ 13same name
    5. yike-gunshi/forge-skills★ 13same name
    6. coreweave/forge-agent-lens-for-claude-code★ 9same name
    7. wandb/forge-agent-lens-for-claude-code★ 9same name
    8. wandb/forge-claude-code★ 9same name
    9. wandb/weave-claude-code★ 9same name
    10. lionshilov/Forge★ 5same name

    1 more not listed

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • Used alongsideProjects with it also use forge-config, forge-install and 3 more.

    Other files people use alongside this one.

    Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
    • coreweaveSkill

      forge-config

      This skill should be used when the user wants to "configure Forge", "set Forge project", "change Forge project", "set wandb api key", "update Forge settings", "show Forge config", "change Forge configuration", "restart the Forge daemon", "apply Forge config changes", "restart Forge to pick up changes", or needs to read or update any Forge Agent Lens for Claude Code plugin settings.

      Also in 4 projects that use this

    • coreweaveSkill

      forge-install

      This skill should be used when the user wants to "install the Forge plugin", "set up Forge", "install forge-agent-lens-for-claude-code", "configure Forge for the first time", "get started with Forge tracing", or needs to complete the initial setup of the Forge Agent Lens for Claude Code plugin including dependency installation and project configuration.

      Also in 4 projects that use this

    • antonailogSkill

      forge-qa

      QA 纯验收（只测不修）：Mode A 完整验收（test-spec → 10 维度断言引擎 → 报告 + User Gate）；Mode B 供 forge-bugfix P6 调用做单 bug 回归并回填 BF 报告。 铁律：每个测试必须有 pass/fail 和深层断言；console.error 零容忍；不猜端口，只用传入的 app_url。 触发方式：Mode A 用户说"测试"、"QA"、"forge-qa"或 forge-dev 调度；Mode B 由 forge-bugfix P6 调用（传 review_doc）。

      Also in 3 projects that use this

    • redteamlifeSkill

      forge-ship

      Use when the user asks to merge, release, promote, reconcile, or close out work in a repo containing `docs/forge/` or governed by FORGE. Not for per-task commits; those complete inside forge-execute-task.

      Also in 3 projects that use this

    • tonmoy007Skill

      forge-review

      Run a multi-dimensional parallel code review over the current change.

      Also in 3 projects that use this

  • In this repository29 other files it gives its agents.

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