pp-lda-gov
Pass
Audited by Gen Agent Trust Hub on Aug 20, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill requires installing an external CLI tool from a specific NPM package (@mvanhorn/printing-press-library) or a Go module (github.com/mvanhorn/printing-press-library). These sources are associated with the skill's author context.
- [COMMAND_EXECUTION]: The skill executes the
lda-gov-pp-clibinary with arguments derived from user input or agent reasoning to fetch and process records. - [DATA_EXFILTRATION]: The CLI tool includes a
--deliverflag that supports sending command output to external URLs via awebhook:<url>sink, providing a mechanism for data exfiltration. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from the LDA.gov API and local SQLite databases, which presents a surface for indirect prompt injection attacks.
- Ingestion points: LDA.gov API filing records, registrant data, and lobbyist information fetched into the agent's context.
- Boundary markers: None specified in the instructions to distinguish between data and potential instructions.
- Capability inventory: The agent has the ability to write output to the local file system or send it to remote webhooks.
- Sanitization: The instructions do not define any sanitization or validation steps for the content retrieved from public lobbying records.
Audit Metadata