orca-per-workspace-env
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill directs the agent to identify and execute local binaries (
orca,orca-dev,orca-ide) or a binary specified in theORCA_CLI_COMMANDenvironment variable. This pattern relies on the integrity of the local environment and path. - [PRIVILEGE_ESCALATION]: The instructions explicitly guide the agent to re-run commands with "escalated permissions" if a
runtime_access_deniederror occurs. This encourages the agent to seek administrative or root privileges to bypass security sandboxes. - [INDIRECT_PROMPT_INJECTION]: The skill loads a "version-matched guide" using the command
ORCA skills get orca-per-workspace-env. The output of this command is intended to drive the agent's subsequent actions, creating an attack surface where a compromised binary could inject malicious instructions into the session. - Ingestion points: The output of the
ORCA skills getcommand inSKILL.mdis loaded as a guide for the agent. - Boundary markers: The skill lacks explicit boundary markers or instructions for the agent to treat the external guide content as untrusted.
- Capability inventory: The skill possesses the capability to execute shell commands and interact with the local filesystem through the
ORCAbinary. - Sanitization: No sanitization or validation of the guide output is mentioned before the agent is expected to follow its instructions.
Audit Metadata