The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to perform deep repository analysis, which includes reading git log, repository source files, and pull request descriptions. This requires the execution of shell commands and interaction with version control tools.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted data from various sources to generate its output.
Ingestion points: In reference/fragments/fragments.md, the agent reads git logs and PR descriptions. In reference/init/init.md, it is directed to fetch and process content from external URLs or local paths provided as writing samples.
Boundary markers: The instructions do not specify any delimiters or safety markers to differentiate between the agent's instructions and the raw material being mined.
Capability inventory: The skill possesses the capability to write to the local file system, including creating new files for articles and modifying global configuration files.
Sanitization: There is no evidence of sanitization or filtering of the ingested content before it is used to influence the agent's writing behavior.
[FILE_SYSTEM_WRITE]: The skill establishes persistence for user preferences by writing to a hardcoded path in the home directory (~/.claude/writing/author.md). While this is a functional requirement for the skill's 'voice profile' feature, it creates a central store of user-specific metadata that could be targeted by other malicious processes.

articulate