The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted entity mentions from document corpora while possessing significant capabilities like shell access and file modification.
Ingestion points: Entity mentions are ingested from external document corpora through the resolve_entities capability and the underlying pipeline described in SKILL.md.
Boundary markers: The instructions do not define clear delimiters or boundary markers to isolate entity data from the agent's instruction context.
Capability inventory: The skill utilizes several powerful tools including Read, Write, Edit, Glob, Grep, and Bash, as specified in the allowed-tools configuration in SKILL.md and schema.yaml.
Sanitization: There is no evidence of sanitization or filtering applied to ingested document content to prevent the execution of embedded instructions.

entity-resolver