The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from external market venues and user-editable diary entries, which serves as an attack surface where malicious instructions could be embedded to influence the agent's trading decisions.
Ingestion points: The skill reads data using tools like jw3_market, jw3_diary, and jw3_analyze as specified in SKILL.md.
Boundary markers: No specific boundary markers or instructions to ignore embedded commands are mentioned in the skill definition to protect against malicious data content.
Capability inventory: The skill possesses significant capabilities to move funds and manage trades via jw3_execute, jw3_close, jw3_exits, and jw3_cancel.
Sanitization: The instructions do not describe any sanitization or validation of the text retrieved from external sources before it is processed by the AI.
[EXTERNAL_DOWNLOADS]: The skill depends on external tools and packages, specifically the @jw3/cli package.
Evidence: The Prerequisites section in SKILL.md explicitly requires the @jw3/cli package and the jw3 CLI tool to be installed on the host system.

jw3-ops