@ssheleg/make-skill 0.20.0 → 0.21.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/CHANGELOG.md CHANGED
@@ -1,5 +1,37 @@
1
1
  # Changelog
2
2
 
3
+ ## v0.21.0 — a gate's self-test runs in both directions
4
+
5
+ *"A validator that can't fail is decoration"* has been the rule here since the standard was
6
+ written. It is **half** the rule, and the missing half is the expensive one.
7
+
8
+ The planted defect proves a gate CAN fire. Nothing in it proves the gate fires only where it
9
+ should — and a gate that flags correct code is switched off within a day, taking the real check
10
+ with it. **A false positive does not arrive as a bug report. It arrives as an argument about
11
+ your gate**, from someone whose code was fine.
12
+
13
+ `references/authoring.md` gains *A gate's self-test runs in both directions*: the two kinds of
14
+ case, what each one's absence costs, and the point that counter-shapes **decide** a gate's
15
+ design rather than merely verifying it — a pattern that finds five hits and is wrong about all
16
+ five is a parser you have not written yet.
17
+
18
+ Two recurring shapes are named because both have shipped:
19
+
20
+ - **The input that cannot be resolved** must fail loudly, never skip — *a check that cannot run
21
+ is not a check that passed*. A slug resolver anchored on `github.com` returned nothing for
22
+ npm's `github:owner/repo` shorthand, so its check printed `skip` and reported success for work
23
+ it never did.
24
+ - **The word inside another word** — `lease` in *please*, `аудит` in *аудитория*, `seo` in
25
+ *Seoul*. Found by running the gate over a real corpus, never by reading it.
26
+
27
+ And: **count the cases, do not restate the count.** A summary saying *"5 cases"* while thirteen
28
+ run is the defect the gate exists to prevent, in the gate's own output.
29
+
30
+ **The doctrine went to the reference rather than the body, and the body's own gate decided
31
+ that.** `SKILL.md` measured 4738 of a 4750 working limit — twelve tokens of headroom — so every
32
+ draft that stated the rule inline failed the budget check. `authoring.md` is linked from
33
+ `SKILL.md` twice, so the rule is reachable without spending a body it does not have. Closes #5.
34
+
3
35
  ## v0.20.0 — the checker said 180 where the file said 1392
4
36
 
5
37
  **Both checkers dropped the continuation lines of a plain multi-line YAML
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@ssheleg/make-skill",
3
- "version": "0.20.0",
3
+ "version": "0.21.0",
4
4
  "description": "Create, retrofit, audit, and ship agent skills & Claude Code plugins the proven ssheleg way — conformance to the Agent Skills open standard AND Anthropic's platform rules (front-matter limits, disclosure budgets, per-surface runtime limits, the Skills API, evals) plus the Claude Code plugin reference (manifest schemas, component layout, claude plugin validate --strict), marketplace repo layout, version sync, validator + CI, multi-channel distribution (plugin, vercel skills CLI, npx, Cursor), npm gotchas, the review checklist for third-party skills, and MCP / A2A rules for protocol-connected skills. This package is the installer CLI.",
5
5
  "keywords": [
6
6
  "skill",
@@ -3,7 +3,7 @@
3
3
  "name": "make-skill",
4
4
  "displayName": "Make Skill",
5
5
  "description": "Create, retrofit, audit, and ship agent skills & Claude Code plugins the proven ssheleg way: conformance to the Agent Skills open standard, Anthropic's platform rules (surfaces, Skills API, evals) and the Claude Code plugin reference, marketplace repo layout, version sync, validator + CI, multi-channel distribution (plugin, vercel skills CLI, npx, Cursor), npm gotchas, end-to-end first publish, the review checklist for third-party skills, plus MCP / A2A references for protocol-connected skills.",
6
- "version": "0.20.0",
6
+ "version": "0.21.0",
7
7
  "author": {
8
8
  "name": "ssheleg",
9
9
  "url": "https://x.com/sshlg93"
@@ -5,7 +5,7 @@ license: MIT
5
5
  compatibility: Authoring works on any agent. The bundled scripts/ need python3. Publishing steps need git, gh, node and npm; the plugin gates need the claude CLI. Not usable on the Claude API surface, which has no network and no runtime package install.
6
6
  metadata:
7
7
  author: ssheleg
8
- version: "0.20.0"
8
+ version: "0.21.0"
9
9
  homepage: https://github.com/ssheleg/make-skill
10
10
  ---
11
11
 
@@ -20,6 +20,7 @@ and [agentskills.io](https://agentskills.io/skill-creation/best-practices).
20
20
  - Workflows and feedback loops
21
21
  - Content guidelines (terminology, time, paths, table of contents)
22
22
  - Scripts — the rules that separate a script from a liability
23
+ - A gate's self-test runs in both directions
23
24
  - Evaluation and iteration — evals before prose
24
25
  - Quality checklist
25
26
 
@@ -187,6 +188,43 @@ run flagged a false positive on the very document that teaches the anti-pattern
187
188
  which is why it now ignores quoted spans. Hooks, subagents and commands that wrap
188
189
  it, plus the fallbacks they owe, are in `references/host-capabilities.md`.
189
190
 
191
+ ## A gate's self-test runs in both directions
192
+
193
+ *"A validator that can't fail is decoration"* is half the rule. The planted defect
194
+ proves the gate **can** fire. Nothing in it proves the gate fires only where it
195
+ should — and a gate that flags correct code is switched off within a day, taking
196
+ the real check with it. **A false positive does not arrive as a bug report. It
197
+ arrives as an argument about your gate**, from someone whose code was fine.
198
+
199
+ So a self-test carries two kinds of case:
200
+
201
+ | Direction | The case | What its absence costs |
202
+ |---|---|---|
203
+ | **Must catch** | the planted defect, watched failing | a check that cannot fail, green forever |
204
+ | **Must NOT flag** | the legal spelling, the empty input, the near-miss that only resembles the defect | a check that is correct and unusable, removed by the first person it wrongly blocks |
205
+
206
+ **The counter-shapes are also where a gate's design gets decided, not just
207
+ verified.** A pattern that reads *"no key may start with a mounted namespace"*
208
+ finds five hits and all five are correct — they pass absolute keys to a translator
209
+ that prefixes nothing. Written as a pattern the gate is wrong; written as a parser
210
+ that binds each translator variable to its namespace, it is right. The counter-case
211
+ is what tells the two apart, and only running it reveals which one you built.
212
+
213
+ Two shapes recur and both belong in every gate's fixtures:
214
+
215
+ - **The input that cannot be resolved.** A resolver returning nothing must be a
216
+ loud failure, never a silent skip — *a check that cannot run is not a check that
217
+ passed*. A slug resolver anchored on `github.com` returned nothing for npm's
218
+ `github:owner/repo` shorthand, so the check printed `skip` and reported success
219
+ for work it never did.
220
+ - **The word inside another word.** Any matcher over prose needs the near-miss:
221
+ `lease` inside *please*, `аудит` inside *аудитория*, `seo` inside *Seoul*. These
222
+ are found by running the gate over a real corpus, not by reading it.
223
+
224
+ **Count the cases, do not restate the count.** A summary line saying *"5 cases"*
225
+ while thirteen run is the same class of defect the gate exists to prevent, in the
226
+ gate's own output.
227
+
190
228
  ## Evaluation and iteration — evals before prose
191
229
 
192
230
  **Build the evaluations before writing extensive documentation.** Otherwise the