@victortomaili/skill-cli 0.6.1 → 0.6.2

package/CHANGELOG.md

@@ -11,6 +11,16 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Cursor adapter (`.cursor/rules` format) for `init -g` bootstrap.
 - Per-agent hook adapters for automatic `/X` triggering (push model).
 
+## [0.6.2] - 2026-07-07
+
+### Changed
+- AGENTS.md START GATE: stricter, broader PROPOSE rule. The agent now auto-loads
+  only skills STRICTLY NECESSARY to attempt the task, and PROPOSES (waits for user
+  confirmation) everything else — explicitly including: response STYLE/format/mode
+  skills, CONTEXT-COST / token optimizers, ACTIVATE-BEFORE-WORK workflows, and any
+  facilitator/accelerator that would make the job EASIER or FASTER. When unsure,
+  propose (propose is the safe default; loading is silent).
+
 ## [0.6.1] - 2026-07-07
 
 ### Changed

package/README.md

@@ -198,13 +198,14 @@ instruction file (`CLAUDE.md`, `AGENTS.md`, `GEMINI.md`):
 
 > **START GATE:** on the first user message, your VERY FIRST action — BEFORE
 > ANYTHING ELSE (before thinking, before any tool call) — is `skill active`, the
-> skill **catalog** (every skill's name + full description; never the body). Read
-> it and decide per skill from its description: **functional** for the task → load
-> it (`skill cat <name>`); **context-altering** (changes HOW you respond —
-> style/mode) → **propose**, apply only after the user confirms (`/X` = confirm).
-> **LOADED ≠ LISTED**: a skill is loaded only if you `cat`-ed it this session —
-> listing it (or its ★ / `active` status) is not loading. Any skill can be
-> context-altering; the agent judges from the description (no flag, no fixed list).
+> skill **catalog** (every active skill's name + full description; never the body).
+> Read it and decide per skill from its description: **strictly necessary** to even
+> attempt the task → load it (`skill cat <name>`); **otherwise PROPOSE** — including
+> any skill that changes your response *style/format*, optimizes *context cost*, is
+> meant to be *activated before the work*, or would make the job *easier/faster*
+> (a facilitator). Apply only after the user confirms (`/X` = confirm); when unsure,
+> propose. **LOADED ≠ LISTED**: a skill is loaded only if you `cat`-ed it this
+> session — listing it (or its ★ / `active` status) is not loading.
 > `/X` → `skill trigger X`.
 
 It's wrapped in `<!-- BEGIN skill-cli --> … <!-- END skill-cli -->` markers, never

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@victortomaili/skill-cli",
-  "version": "0.6.1",
+  "version": "0.6.2",
   "description": "Cross-agent skill manager — one global store, clean agent folders, activation via skill.config",
   "license": "MIT",
   "author": "Victor Tomaili <victor@tomaili.com>",

package/src/commands/defaults.js

@@ -16,7 +16,7 @@ export function cmdActive() {
   const eff = computeEffective(installed, globalCfg, projCfg)
   const defs = new Set((globalCfg.defaults || []).map(d => String(d).toLowerCase()))
 
-  console.log(c.bold('skill active') + c.gray(' — active skills in this project (descriptions only; functional → `skill cat`, context-altering → propose)'))
+  console.log(c.bold('skill active') + c.gray(' — active skills (descriptions only). Load only what is STRICTLY necessary; PROPOSE style/cost/workflow/facilitator skills and wait for confirmation.'))
   console.log()
 
   if (eff.length === 0) {

package/src/lib/agents-md.js

@@ -28,10 +28,19 @@ action — BEFORE ANYTHING ELSE: before you start thinking/reasoning about the t
 and before any tool call — MUST be to run \`skill active\`. It prints the skill
 CATALOG — every ACTIVE skill's name + FULL description (NOT the skill body).
 Read it, then for EACH skill decide FROM ITS DESCRIPTION:
-- Functional / useful for THIS task → load it: \`skill cat <name>\`.
-- Changes HOW you respond (a style, tone, format, or persistent MODE — e.g. a
-  brevity/compression style) → it is CONTEXT-ALTERING: PROPOSE it to the user and
-  apply ONLY after they confirm (\`/X\` = confirm). Do NOT load/apply it yourself.
+- It is STRICTLY NECESSARY to even attempt the task (you cannot do it without
+  this knowledge) → load it: \`skill cat <name>\`.
+- OTHERWISE PROPOSE it — tell the user it's available and what it does, then apply
+  ONLY after they confirm (\`/X\` = confirm). You MUST propose (never silently apply)
+  any skill that:
+  • changes your communication / response STYLE, tone, format, or a persistent
+    mode (e.g. a brevity / compression style);
+  • optimizes CONTEXT COST / token usage;
+  • is meant to be ACTIVATED BEFORE the work (a workflow / methodology);
+  • would make the job EASIER or FASTER — a facilitator / accelerator / tool —
+    even if optional. Surface it so the user can opt in.
+When unsure whether a skill is necessary or optional → PROPOSE (proposing is the
+  safe default; loading is silent).
 LOADED ≠ LISTED. A skill is loaded ONLY if you have run \`skill cat <name>\` for
 it THIS session. Seeing a skill in \`skill active\` / \`skill list\` does NOT load
 it — you then know only its name + description, nothing more. Never infer a skill