菜单
Use when the user asks to forge, scaffold, or author a new cartridge — content cartridge, department cartridge, or chipset bundle. Drives the cartridge-forge loop (scaffold → fill → gate checks → commit) via the skill-creator cartridge CLI.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/Tibsfox/gsd-skill-creator --skill cartridge-forge命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
基于 SOC 职业分类
Use the moment a request is underspecified and you are about to act on an ASSUMED goal. Enumerate the candidate goals the request could mean, project the action or artifact each one implies, and check whether they diverge: if two or more plausible goals lead to materially different actions or artifacts, ask exactly ONE targeted clarifying question; otherwise proceed on the most-likely goal and state the assumption in one line. Clarification is an evidence-producing action, not a delay — but one question, never an interrogation. Distinct from gsd-spec-phase (GSD-phase-bound, emits a heavy SPEC.md) and intent-router (fetch strategy, not goal disambiguation). Backed by agent goal-state inference (arxiv 2606.16813v1). Triggers on acting under an assumed goal when the request admits more than one materially different reading.
Use at write time to vector/embedding memory — Grove content-addressed insertion, chroma/pgvector upserts, memory-consolidation promoting session traces to MEMORY.md, or embedding externally-ingested content. Scores each candidate record against a fixed panel of sentinel queries and quarantines any record that would become the nearest neighbor of too many unrelated queries — a hub — whether from adversarial poisoning or accidental over-generality. This is the memory-record-side sibling of skill-injection-guardian (file-side) and the write-side complement of memory-use-warrant (read-side). Quarantine, never silently drop; a human reviews. Backed by the admission-time hubness gate (arxiv 2606.19692v1). Triggers on inserting into vector memory, consolidating memory, and embedding stranger content.
Run this appropriateness check the moment you are about to integrate a retrieved long-term memory — a Grove content-addressed hit, a chroma/pgvector neighbour, a memory-consolidation digest, or a MEMORY.md line — into a response, especially anything touching private origins, Fox Companies IP, credentials, or Center Camp / consent-governed content. It answers a question intent-router never asks: not WHAT to fetch or HOW, but WHETHER a correctly-retrieved item should reach output. Relevance is not appropriateness — a perfect similarity match can still be a boundary violation. Default is FAIL-CLOSED: if in-context authorization is uncertain, the memory may inform behaviour but must not be surfaced. Backed by RBI-Eval (arxiv 2606.06055v1). Triggers on surfacing recalled sensitive memory into a response.
正在显示 SKILL.md
| name | cartridge-forge |
| description | Use when the user asks to forge, scaffold, or author a new cartridge — content cartridge, department cartridge, or chipset bundle. Drives the cartridge-forge loop (scaffold → fill → gate checks → commit) via the skill-creator cartridge CLI. |
| description-frequency | on-demand |
| user-invocable | true |
| type | skill |
| category | gsd |
| status | stable |
| origin | tibsfox |
| modified | false |
| first_seen | "2026-04-15T00:00:00.000Z" |
| first_path | examples/skills/gsd/cartridge-forge/SKILL.md |
| superseded_by | null |
Forge a new cartridge end-to-end using the skill-creator cartridge
CLI. The forge is the toolchain that builds cartridges; this skill
is the playbook that drives it.
Trigger on any of:
Do not trigger for migrating legacy cartridges (that is
cartridge migrate) or for forking an existing one (that is
cartridge fork).
docs/cartridge/FORGING-GUIDE.md and
docs/cartridge/CARTRIDGE-SPEC.md are the normative references.
Read them once per session if the context is not already loaded.examples/cartridges/<slug>/ for in-repo cartridges./my-cartridges/<slug>/ for ad-hoc / user-local work-department suffix unless the
cartridge is directly tied to a college department in
.college/. "systems-administration" — not
"systems-administration-department".department (skills + agents + teams),
content (deepMap + story arc), coprocessor (pure functional
tooling), or graphics (GLSL / OpenGL / WebGL / Vulkan shader
pipeline; grounded in the GFX research series at
/Research/GFX/ on tibsfox.com).scaffold → fill → validate → eval → dedup → metrics → commit
Every step is idempotent and JSON-capable. eval is the gate —
if it exits non-zero, the cartridge is not ready to ship.
skill-creator cartridge scaffold <template> <dir> <slug>
Writes an 8-file skeleton: cartridge.yaml, chipsets/*.yaml,
one placeholder skill, agent, and team, plus a README.md.
Round-trips through load → validate with zero errors before
any edits.
Replace placeholder content with real content. A full department cartridge typically wants:
is_capcom: true on the capcomuse_whendomain: field on each skillReplace the skeleton's cartridge.yaml header:
id: slugname: title case, human-readableauthor: real authordescription: one-paragraph, no marketing toneprovenance.origin, provenance.createdAtmetadata.tagsDelete the three placeholder markdown files
(skills/placeholder-skill.md, agents/placeholder-agent.md,
teams/placeholder-team.md) and replace them with one markdown
file per real skill / agent / team. Each companion file is
informational — the loader does not scan them, but a "full"
cartridge ships them.
skill-creator cartridge validate <dir>/cartridge.yaml --json
valid: true is the required outcome. If there are validation-debt
errors (e.g. around agent_affinity or domains_covered), add
--allow-validation-debt and track the debt per
docs/cartridge/KNOWN-VALIDATION-DEBT.md.
skill-creator cartridge eval <dir>/cartridge.yaml
Runs the pre-deploy gates declared in
chipsets/evaluation.yaml. Default gates:
all_skills_have_descriptionsall_agents_have_rolesgrove_record_types_definedhas_evaluation_chipsetExit code 0 is required before commit.
skill-creator cartridge dedup <dir>/cartridge.yaml
no collisions is required. Collisions mean a skill or agent is
defined twice across chipsets — either intentional merge
(refactor) or typo (fix).
skill-creator cartridge metrics <dir>/cartridge.yaml
Informational, not a gate. Captures shape (skill / agent / team / record-type counts) for the commit message.
Conventional Commits, scope cartridge:
feat(cartridge): forge <slug> cartridge
N skills / M agents / K teams / J grove record types.
Forged via cartridge-forge from the <template> template.
All forge gates green: validate, eval, dedup.
cartridge.yaml from scratch. Always
scaffold first — the template encodes the current schema.eval. It is the ship gate.kind: department is a schema discriminator, not a college
tie. The chipset kind stays department even when the
cartridge is not tied to a college department.-department from the slug unless the cartridge is
tied to an entry under .college/.<slug>-dr-drill-team or equivalent) if the cartridge has
data-durability responsibilities.A complete worked example:
examples/cartridges/systems-administration/ — 12 skills,
5 agents, 4 teams, 6 grove record types, all forge gates green.
After the loop completes, report to the user:
id, name, and pathvalidate / eval / dedup verdictsDo not report "done" before all four forge gates (validate, eval, dedup, metrics) have been run and shown green.