qyl-mcp-server 8.1.0 → 8.1.1
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/README.md +5 -0
- package/dist/mcp-app.html +1 -1
- package/dist/mcp-dashboard.html +1 -1
- package/dist/mcp-home.html +1 -1
- package/dist/skills/qyl-investigate/SKILL.md +1 -1
- package/package.json +1 -1
package/README.md
CHANGED
|
@@ -198,6 +198,11 @@ all 32 tools against an isolated collector started by the published CLI.
|
|
|
198
198
|
|
|
199
199
|
## Release notes
|
|
200
200
|
|
|
201
|
+
### 8.1.1
|
|
202
|
+
|
|
203
|
+
- Package the corrected qyl-investigate skill: `verify_fix` needs `since` or `release` beside the `failure_id`.
|
|
204
|
+
- Documentation only otherwise: the plugin bundle README, the Anthropic listing draft and both plugin manifests describe the 32-tool surface; the release contract records the OpenAI plugin draft and the verified domain; the threat model states the single scope check. No code change since 8.1.0.
|
|
205
|
+
|
|
201
206
|
### 8.1.0
|
|
202
207
|
|
|
203
208
|
- Serve the qyl-investigate skill from the server through the MCP skills extension (`io.modelcontextprotocol/skills`: `skills/list`, `skills/get` and `resources/read` under `skill://qyl/`), so the OpenAI plugin portal imports it at Scan Tools and the npm package carries it in `dist/skills`. The skill now names the tool to start with for each question, the defaults every tool applies, the accepted `aggregate_telemetry` measures, and the one-call connection rule: no fan-out before one call has returned telemetry, one retry after sign-in, then stop.
|