The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external and potentially untrusted brand content (e.g., website copy, PDF guidelines, chat exports) in Phase 1A to distill a brand persona. This creates a surface for indirect instructions to be embedded in the source data.
Ingestion points: External brand documents, chat exports, and social media archives provided during Phase 1A (SKILL.md).
Boundary markers: The skill does not define explicit delimiters or instructions to ignore embedded commands within the ingested brand content.
Capability inventory: The skill has access to Bash, Write, Read, and WebSearch tools (SKILL.md).
Sanitization: There is no evidence of sanitization, filtering, or validation of the content extracted from external documents before it is used in the distillation process.
[COMMAND_EXECUTION]: The skill performs shell command execution to manage the persona generation lifecycle and network registration.
Evidence: SKILL.md Phase 4a executes npx openpersona create --config persona.json --output ./{slug}-skill to generate the final persona pack.
Evidence: SKILL.md Phase 5 executes npx openpersona acn-register {slug} --endpoint <url> to register the agent on the Agent Communication Network.
[EXTERNAL_DOWNLOADS]: The skill facilitates the installation and runtime execution of external code from public registries.
Evidence: The skill facilitates the installation and execution of the openpersona package via npx or npm install (README.md).

brand-persona-skill