The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses the bunx utility to download and run the @rayhanadev/truffler package from the NPM registry at runtime.
[REMOTE_CODE_EXECUTION]: By invoking bunx @rayhanadev/truffler, the skill enables the execution of remote code fetched during the rule-writing process.
[COMMAND_EXECUTION]: The instructions guide the agent to execute repository commands such as nr test and nr lint. These commands run arbitrary scripts defined in the local package.json file.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes external rule-research contracts and user-provided instructions to generate code logic without explicitly defined safety boundaries.
Ingestion points: The rule-research contract and user implementation requests described in the detector planning section.
Boundary markers: The skill lacks explicit instructions or delimiters to isolate untrusted contract data from the agent's core instructions.
Capability inventory: The skill possesses the ability to execute shell commands (bunx, nr) and modify files in the repository.
Sanitization: No sanitization or validation logic is specified for the data ingested from external contracts.

rule-writing