The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute a local Python script (ledger.py) to record, differentiate, and trend SEO metrics. The command incorporates a user-provided <domain> argument, which is a common pattern for CLI-based skills but involves subprocess spawning.
Evidence: Found in SKILL.md under the 'Zero-dependency measurement loop' section: python3 "${CLAUDE_PLUGIN_ROOT}/scripts/connectors/ledger.py" record <domain> --source monitor --data '{"sessions": ..., "clicks": ..., ...}'.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from various external SEO and marketing sources (Google Search Console, Analytics, and third-party SEO tools) to generate dashboards and alerts, creating a potential surface for indirect prompt injection.
Ingestion points: Data is pulled from analytics exports, Search Console inputs, and external SEO monitor tools as described in SKILL.md.
Boundary markers: The provided output templates in references/report-output-templates.md and references/report-templates.md do not include explicit delimiters or 'ignore instructions' warnings for external data.
Capability inventory: The skill has the capability to write performance reports and alert configurations to the local file system (memory/monitoring/).
Sanitization: There is no documented logic for sanitizing or validating the content of the ingested metrics or competitor data before interpolation into the final report or memory storage.

performance-monitor