The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill possesses an indirect prompt injection attack surface. It instructs the AI agent to read and process untrusted data from the local execution environment (specifically __manifest__.py files) and from external GitHub repositories (Odoo and OCA) to gather development context.
Ingestion points: The main entry point SKILL.md and the agents/odoo-context-gatherer.md workflow specifically direct the agent to read local project manifest files and search remote Odoo Community Association (OCA) repositories.
Boundary markers: The skill does not provide explicit delimiters or instructions to the agent to treat this ingested content as data only and to ignore any natural language instructions that might be embedded within them.
Capability inventory: The skill grants the agent extensive capabilities, including creating/modifying Python and JavaScript code, executing shell commands for file searching (e.g., grep), and performing automated code reviews.
Sanitization: No sanitization or filtering logic is specified for the external data being processed.
[SAFE]: Aside from the inherent risk of data ingestion, the code patterns and architectural guidelines provided are highly secure and follow Odoo development standards. For example, the skill provides secure templates for webhook handling using HMAC signature verification, encourages the use of Odoo’s configuration parameter system instead of hardcoded credentials, and includes explicit warnings to developers about the risks of using the sudo() method to bypass security rules.

odoo-development-skill