The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's core function involves processing untrusted data captured from external websites, which creates a vulnerability to indirect prompt injection.
Ingestion points: External data is ingested into the system via hooks/memory-palace-index.yaml and stored markdown files generated from WebFetch and WebSearch actions.
Boundary markers: While the skill documentation mentions that capture hooks escape YAML scalars to prevent parsing failures, it does not specify the use of delimiters or explicit instructions to the AI agent to ignore instructions embedded within the captured content when it is surfaced during a session.
Capability inventory: The skill possesses capabilities to execute local scripts (uv run python), perform file system modifications (updating the index and frontmatter), and create backups in data/backups/.
Sanitization: Content is escaped for YAML compatibility at the ingestion stage, but there is no evidence of semantic sanitization to filter out malicious LLM instructions before the data is presented to the agent.
[COMMAND_EXECUTION]: The skill workflow requires the execution of multiple local scripts to perform corpus maintenance and indexing.
Evidence: The instructions direct the agent to execute scripts/memory_palace_cli.py, scripts/build_indexes.py, and scripts/repair_capture_frontmatter.py to manage the index.
Evidence: The skill documentation notes the automated execution of scripts/precommit_palace_maintenance.sh as part of the git workflow to ensure the capture backlog is processed before commits.

palace-index-curator