The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external brand profiles and guideline files which could serve as a vector for indirect instructions.
Ingestion points: The skill reads from ~/.claude-marketing/brands/{slug}/profile.json and ~/.claude-marketing/brands/{slug}/guidelines/_manifest.json as specified in SKILL.md.
Boundary markers: There are no explicit delimiters or instructions provided to the agent to disregard potential AI instructions embedded within these JSON or markdown files.
Capability inventory: The skill has capabilities for file system access (read) and local command execution via Python scripts.
Sanitization: No sanitization or validation of the data content is described before it is used in the prompt context.
[COMMAND_EXECUTION]: The skill executes local Python scripts to manage campaign and benchmark data.
SKILL.md instructs the agent to run python "${CLAUDE_PLUGIN_ROOT}/scripts/campaign-tracker.py" --brand {slug} --action list-campaigns to retrieve history.
affiliate-marketing.md references a requirement to use python scripts/benchmark_book.py --action record ... --source <url> to maintain updated market figures.
The use of the {slug} parameter in a shell command represents a surface for potential argument injection if the brand identifier is not properly validated.

growth-engineering