manuscript-manager Eval report.

Validation plan

What the agent should do next

1. 1Inspect repository, README/SKILL.md, license, and recent commits before production use.
2. 2Install in an isolated workspace or sandbox with no production secrets available.
3. 3Run the smallest representative task and record files touched, commands run, network access, and outputs.
4. 4Compare the selected skill against at least one alternative when the eval status is review or failed.
5. 5Promote only after the agent reports a successful verification result and unresolved warnings are accepted.

Do not use when

Conditions that require another skill

• teams that need a vendor-supported SLA
• production agents without a repository review
• SKILL.md does not explicitly instruct the agent to treat PROJECT_STATE.yaml and other project files as untrusted data, leaving room for prompt injection if a state file contains instructions.
• Safe operating boundaries are not fully explicit: the skill mentions creating missing folders and updating PROJECT_STATE.yaml, but does not clearly prohibit destructive actions like deleting or rewriting chapter files.
• The consistency check for word count and chapter status is underspecified; different agents may count words or infer status differently.
• Quality score needs review