The Agent Skills Directory

[COMMAND_EXECUTION] (MEDIUM): The skill requires the agent to execute the brv binary on the host system to perform query and curation tasks. This relies on an external, unverified dependency being present in the environment.
[DATA_EXFILTRATION] (HIGH): The brv curate command explicitly reads local files (using the -f flag) and sends their content to a 'Context Tree' managed by ByteRover Inc. (byterover.dev). While this is the primary purpose of the skill, it facilitates the bulk transfer of source code and potentially sensitive configuration to an external third-party service. Severity is reduced to MEDIUM as this is the primary stated purpose.
[INDIRECT_PROMPT_INJECTION] (LOW): The skill is vulnerable to indirect prompt injection through its core data loop.
Ingestion points: Data enters the agent context through brv query results and file contents read via the -f flag in SKILL.md.
Boundary markers: No explicit boundary markers or instructions to ignore embedded commands within the retrieved context are provided.
Capability inventory: The skill possesses shell execution capabilities (executing brv) and file read capabilities (passing file paths to brv).
Sanitization: There is no evidence of sanitization or validation of the data retrieved from the 'Context Tree' before it is processed by the agent.

byterover