The Agent Skills Directory

[SAFE]: The skill is designed for informational interpretation of logs and result files located in the specific application directory ~/.claude-octopus/results/. It focuses on tallying and summarizing data without performing network operations or accessing sensitive system credentials.
[SAFE]: The skill includes explicit instructions for the agent to avoid automatic execution of decisions, stating that verdicts gate decisions but do not execute them. This serves as a procedural guardrail against unintended actions.
[SAFE]: Frontmatter configuration disable-model-invocation: true further restricts the skill's ability to trigger external model actions, limiting its execution environment to basic data processing and reporting.

council-verdicts