neo-agent-skills 0.1.26 → 0.1.28

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.
@@ -147,15 +147,13 @@ Map vs World Atlas applies **recursively**. A workflow file (`references/<workfl
147
147
 
148
148
  ## 3. The Lesson Promotion Path
149
149
 
150
- When a swarm agent discovers a systemic trap, an architectural pattern, or a workflow optimization that took significant effort to derive, that knowledge must not die when the session ends.
150
+ Promote systemic lessons through the existing skill that governs the failed decision.
151
151
 
152
- **You MUST promote valuable operational lessons to the Swarm:**
152
+ 1. Keep the evidence, owner and disposition on the existing source artifact; apply the correction or reject it with evidence.
153
+ 2. Express runtime behavior as `Bias`, `Rule`, `Rationale`, `Trigger`. Keep incident history and examples behind provenance pointers.
154
+ 3. Rewrite the smallest trigger or payload that changes the decision. Create a skill only for a new operational domain.
153
155
 
154
- 1. **Locate the relevant domain:** Determine which existing skill governs the domain (e.g., `pull-request`, `neural-link`, `unit-test`).
155
- 2. **Classify before writing:** Runtime behavior becomes a compact decision atom: `Bias`, `Rule`, `Rationale`, `Trigger`. Incident history, examples, and provenance move behind atlas/provenance pointers instead of entering the runtime path.
156
- 3. **Update the smallest surface:** Edit the existing payload section that fires for the domain. Author a new skill only when the lesson represents a genuinely new operational domain.
157
-
158
- *Why:* Skills are the permanent architectural memory of the swarm. Promoting lessons ensures the next agent does not repeat your expensive mistakes.
156
+ A deferred lesson retains an owner and an observable activation condition beyond its parent's closure. Record the shipped source and the consuming version's load receipt; validate the behavior on the next named triggering case. Saved, shipped, loaded and validated are different states. A recurrence reopens the correction, not another copy of the retrospective.
159
157
 
160
158
  ## 4. The Claude Symlink Mandate
161
159
 
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "neo-agent-skills",
3
- "version": "0.1.26",
3
+ "version": "0.1.28",
4
4
  "description": "Install the workflow substrate a cross-model AI team actually runs on: ticket intake, PR review gates, lane coordination and guards, versioned as a package. The skills Neo.mjs maintainers load every session — materialized into your repo, not copied.",
5
5
  "type": "module",
6
6
  "license": "MIT",