The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes multiple subcommands of the skillwiki CLI, including path, health, lint, and log-rotate. These commands are used to analyze vault data and perform maintenance tasks. The log-rotate and lint --fix commands are gated by user consent or specific security exceptions.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it instructs the agent to read and reason over the output of vault health and lint reports, which are derived from potentially untrusted user content within the vault.
Ingestion points: Data is ingested from the output of skillwiki health <vault> --out /tmp/skillwiki-health.json and skillwiki lint <vault> --summary in SKILL.md.
Boundary markers: The instructions do not define specific boundary markers or delimiters for the agent to use when parsing the JSON envelope or lint summaries.
Capability inventory: The agent has the ability to execute shell commands via the skillwiki tool, including commands that modify the filesystem like log-rotate --apply and lint --fix --only sensitive_content.
Sanitization: No sanitization or instruction-filtering is applied to the data retrieved from the vault before the agent performs reasoning (Step 4).

wiki-lint