tk-studio-install
Install the BMad base at the committed bmad.lock pin with the project's module set — non-interactive, deterministic, zero forks. Use when the user says "tk install", "install the base", or "install bmad at the pin".
- License 许可
- MIT
- 最近更新
- 2026-11-10
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- tk-studio-install
- description
- Install the BMad base at the committed bmad.lock pin with the project's module set — non-interactive, deterministic, zero forks. Use when the user says "tk install", "install the base", or "install bmad at the pin".
02 资源文件 2
- 其他customize.toml470 B
- 脚本scripts/install_base.py9.8 KB
03 SKILL.md 正文
tk-studio-install
Installs the BMad base into a project at exactly the versions bmad.lock pins
(AD-1). The upstream installer does all the work — this skill only renders the
pinned invocation, verifies the result, and measures the outcome.
Behavior (both modes)
-
Run the installer script through the plugin root:
uv run "${CLAUDE_PLUGIN_ROOT}/skills/tk-studio-install/scripts/install_base.py" --directory <project-root>Optional:
--modules a,b,cto install a subset of the lock's module set (the project's configured module set once project config exists, Epic 2);--dry-runto show the exactnpxcommand without running it;--no-normalizeto keep the upstream installer's output byte-for-byte. -
The script is the whole flow: parse lock → run
npx bmad-method@<pin> installnon-interactively (--yes, per-module--pinflags, stdin closed) → verify_bmad/_config/manifest.yamlmatches every pin → emit oneinstall-outcomeevent (success, or failure with the failing step) → normalize the reinstall's no-op churn. Trust its JSON output; do not re-run the installer by hand.Normalization (EP-011 D2, shared with
tk-studio-base-update) runs only after verify-at-pin passes, and only when the project root is a git top-level with_bmad/tracked. An untracked or ignored_bmad/has no committed tree to be a no-op against. It restores files upstream rewrote to something provably equivalent to the committed blob — line endings, list values re-serialized as JSON strings, manifestlastUpdatedstamps, derivativefiles-manifest.csvhashes — and drops*.bakcopies identical to the committed file. Anything already dirty before the run is left alone. The result'snormalizedobject carries per-class counts andremaining: entries that still differ from HEAD because upstream changed them for real.Base patches (0.1.17) are re-applied before normalization: the studio's declared fixes to installer-owned files (
base-patches/patches.json,lib/basepatch.py), each pinned to one upstream issue, the exact upstream text it replaces, and the core versions it was verified at. A reinstall therefore never silently reverts one. The result'sbase_patcheslists each patch's state. -
Report:
-
Attended: state the installed core/module versions, or on failure the failing step (
upstream-installervsverify-at-pin) and its detail. Whennormalized.remainingis 0 the install was a no-op against the committed tree — say so. When it is above 0, the installer wrote real changes at this pin: in the studio repo those land throughtk-studio-base-update(a same-version re-affirmation), never a hand commit; in any other project they are the project's to review and commit.normalized.skippednames why normalization did not run. Name every base patch whose state isstaleorunverified-corewith its upstream issue: the upstream file changed or the core moved, so the patch was NOT applied and must be re-verified (retire it if the issue is fixed at the pin). -
Headless: no prompts (AD-11). End with the status block —
completeon exit 0;blockedwith the failing step asreasonotherwise:{"status": "complete", "intent": "tk-studio-install", "artifacts": ["_bmad/"], "reason": null}
-
If the deterministic core is unrunnable — a tool call denied by permissions, uv/python unavailable — end blocked with the status block naming the unrunnable core as reason: never a question, never a headless run that ends without the block (AD-11).
Rules
- Never edit anything under
_bmad/— upstream owns that tree (NFR1); a bad install is fixed by rerunning at the pin, never by patching. Normalization is not an edit: it only restores committed bytes upstream itself wrote. - Never bump a pin here — pin changes are
tk-studio-base-update's job alone. - A long installer run is normal (npx cold cache); the script caps it with
--timeout(default 900s) rather than hanging forever.
04 同仓库更多 Skills
tk-studio-track
Tracking verbs on the bound Jira project — `block` moves a story the loop cannot progress, and each dependent story, to "Blocked - Awaiting User" with a comment that @mentions the owner (or, with `--engagement`, one engagement's questions issue, which `close` later moves to "Done"); `reply` posts the owner's answer (or "Done.") with the owner's host-only token and moves the issue to "Block Resolved"; `validate` checks that resolution and passes the issue back to work or re-blocks it with a follow-up. Use when the user says "block this story", "tk track block", "put it in the owner's inbox", "reply to the blocked issue", "validate the resolution", or a supervisor or engagement payload names this skill.
tk-studio-session
Session discipline as a surface — land a boundary handoff (optionally carrying knowledge deltas) and end the session, or resume a fresh session from a run workspace alone. Use when the user says "hand off this run", "tk session", "resume the run", or a boundary/budget trigger says the session must split.
tk-studio-research
Run a chartered research pass — survey the topics a charter scopes, grade every finding by evidence, land the findings in the run workspace, and emit observation events for recommendation-worthy ones. Use when the user says "run the ecosystem watch", "tk research", or a research job's invoke-skill directive names this skill.
tk-studio-report
One-verb human defect report — "the skill did the wrong thing" — recorded as a report event in the local measurement ledger. Use when the user says "tk report", "report a defect", "that skill misbehaved", or "log a studio issue".
tk-studio-playable-judge
The playable-verify leg's screenshot judge (Story 34.3) — reads only a playable run's report.json and its frames, answers the fixed rubric (scene loaded, player visible and framed, no magenta or error overlay, HUD reads, the named moments shown) with pass or fail and findings each naming a frame and a rubric line. Driven headless by the supervisor's host leg on the project's sandbox tier; use when the payload names a playable `media` folder, or the user says "judge the playable frames".
tk-studio-plan-sync
Planning adapter — canonicalize BMad-native planning artifacts (normalize pass, sole id authority) and sync them with the bound planning backend. Use when the user says "sync the plan", "normalize planning artifacts", "tk plan sync", or after stock BMad planning flows produce or change epics/stories.