The Agent Skills Directory

[SAFE]: The skill is a reasoning strategy guide focused on architectural and process improvements. It does not contain any executable scripts or dangerous commands.
[EXTERNAL_DOWNLOADS]: The skill references a GitHub repository (github.com/jacob-balslev/skill-graph) belonging to the skill's author. This is a legitimate vendor resource for metadata synchronization.
[INDIRECT_PROMPT_INJECTION]: While the skill instructions describe processing external data like session logs and code audits, it lacks capabilities such as network operations or file writing that would be necessary to exploit an injection. It focuses on normalization and hashing for clustering, which is a safe way to handle untrusted input.
[COMMAND_EXECUTION]: The skill includes examples of grep commands for auditing codebases. These are illustrative examples intended for use with the agent's standard Grep tool and do not involve privilege escalation or dangerous side effects.

pattern-recognition