Version Packages

packages/cli/CHANGELOG.md

@@ -1,5 +1,34 @@
 # @gtbuchanan/cli
 
+## 0.7.1
+
+### Patch Changes
+
+- 23ae7e2: Generate `deploy:skills` for skills at the workspace root
+
+  A monorepo whose `skills/` sits at the root got no `deploy:skills` task
+  at all: the capability was read from the packages alone, so the root's
+  copy — which `capabilitiesFor` already detects — went unused. The task
+  disappeared the moment `packages:` globs were added to
+  `pnpm-workspace.yaml` and the root stopped being the lone package.
+
+  `gtb sync` now emits a `//#deploy:skills` root task and the root
+  `deploy:skills` script backing it whenever a monorepo root carries its
+  own `skills/`, with `build` depending on it — the shape
+  `//#lint:eslint` already uses. `gtb verify` covers it for free, since a
+  generated task is one it can report as missing.
+
+  Root skills are not a fallback for per-package ones: both are emitted
+  when both exist. They are the case a package can't serve — skills
+  consumed as `<root>/skills/<name>/SKILL.md` are invisible inside a
+  package, and a private workspace package reaches no consumer at all.
+
+  Because turbo dispatches `//#deploy:skills` through the root script of
+  that name, such a root no longer gets the `gtb turbo run deploy:skills`
+  alias — it would re-enter turbo and trip the
+  `recursive_turbo_invocations` guard. Reach every package's copy through
+  `pnpm build` or `gtb turbo run deploy:skills`.
+
 ## 0.7.0
 
 ### Minor Changes

packages/cli/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@gtbuchanan/cli",
-  "version": "0.7.0",
+  "version": "0.7.1",
   "description": "Shared build CLI",
   "type": "module",
   "imports": {

packages/eslint-config/CHANGELOG.md

@@ -1,5 +1,12 @@
 # @gtbuchanan/eslint-config
 
+## 0.5.2
+
+### Patch Changes
+
+- Updated dependencies [79c96d0]
+  - @gtbuchanan/eslint-plugin-agent-skills@1.0.0
+
 ## 0.5.1
 
 ### Patch Changes

packages/eslint-config/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@gtbuchanan/eslint-config",
-  "version": "0.5.1",
+  "version": "0.5.2",
   "description": "Shared ESLint configuration",
   "type": "module",
   "imports": {

packages/eslint-plugin-agent-skills/CHANGELOG.md

@@ -1,5 +1,36 @@
 # @gtbuchanan/eslint-plugin-agent-skills
 
+## 1.0.0
+
+### Major Changes
+
+- 79c96d0: Gate skill length on tokens rather than lines
+
+  `configs.recommended` now caps both `SKILL.md` and
+  `**/skills/*/references/**/*.md` with `agent-skills/max-tokens` at 5000,
+  and applies `agent-skills/max-lines` to neither. The rule still ships, so
+  a repo wanting the spec's 500-line figure can wire it explicitly:
+
+  ```js
+  {
+    files: ['**/skills/*/SKILL.md'],
+    rules: { 'agent-skills/max-lines': ['warn', { max: 500 }] },
+  }
+  ```
+
+  Both rules proxy for how much context a skill costs to load, and tokens
+  measure it directly. At the ~10-14 tokens per line typical of prose
+  skills, 500 lines works out to 5200-7000 tokens, so the token cap is the
+  one that fires. The line cap leads only below ~10 tokens per line, where
+  it reports formatting — semantic line breaks and dense lists raise a line
+  count without changing what the agent loads.
+
+  The 300-line cap on `references/` had no upstream basis: the spec's third
+  tier is "Resources (as needed)" and `skill-creator` calls bundled
+  resources "unlimited, loaded as needed". The 5000 there is a backstop for
+  a reference file costing more to load than the instructions tier it was
+  split out of.
+
 ## 0.3.0
 
 ### Minor Changes

packages/eslint-plugin-agent-skills/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@gtbuchanan/eslint-plugin-agent-skills",
-  "version": "0.3.0",
+  "version": "1.0.0",
   "description": "ESLint plugin and JSON Schema for Agent Skills",
   "type": "module",
   "imports": {