The Agent Skills Directory

[PERSISTENCE]: The skill's onboarding documentation (references/onboarding.md) provides specific commands for the agent to present to the user that append export statements to shell initialization files such as ~/.zshrc and ~/.bashrc. These commands are designed to persist the LINKFOX_AGENT_API_KEY across different terminal sessions. While this is a standard configuration method for CLI tools, it utilizes a persistence mechanism that modifies user shell profiles.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection by ingesting user-supplied keywords and marketplace codes.
Ingestion points: User-provided parameters defined in SKILL.md and passed to scripts/junglescout_keyword_history.py.
Boundary markers: The instructions lack delimiters or explicit warnings for the agent to ignore potentially malicious instructions embedded within the keywords.
Capability inventory: The skill performs network operations via urllib.request to a remote gateway and writes data and cache files to the local project directory.
Sanitization: There is no evidence of input validation or escaping for the user-supplied data before it is processed and transmitted.
[COMMAND_EXECUTION]: The skill requires the execution of local Python scripts (scripts/junglescout_keyword_history.py and scripts/onboarding.py) to perform its primary functions. These scripts execute shell-accessible operations, including retrieving environment variables, managing local file systems for caching, and conducting authenticated network communication with vendor APIs.

linkfox-junglescout-keyword-history