hf-cloud-aws-context-discovery
Warn
Audited by Gen Agent Trust Hub on Aug 15, 2026
Risk Level: MEDIUMCREDENTIALS_UNSAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [CREDENTIALS_UNSAFE]: The skill directs the agent to read the user's
~/.aws/credentialsfile to resolve profiles and configuration. Directly accessing raw credential files exposes sensitive authentication material to the agent session. - [COMMAND_EXECUTION]: The skill uses shell commands, including
aws configure listandaws sts get-caller-identity, to validate credentials and retrieve caller information. These commands provide the agent with deep visibility into the user's cloud identity and permissions. - [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting data from external sources (local configuration files and environment variables) without sanitization.
- Ingestion points:
~/.aws/config,~/.aws/credentials, and environment variables likeAWS_PROFILEandAWS_REGION(SKILL.md). - Boundary markers: Absent; there are no instructions to ignore potential instructions embedded within the configuration values.
- Capability inventory: The agent has the ability to execute
awsCLI commands via subprocess calls. - Sanitization: Absent; the skill does not specify any filtering or validation for the data retrieved from the AWS configuration.
Audit Metadata