angular-developer
Warn
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: MEDIUMMETADATA_POISONINGINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [METADATA_POISONING]: The skill metadata contains deceptive author and copyright information ('Copyright 2026 Google LLC'). This contradicts the provided author context ('l-x-t') and uses a future date, which may mislead users about the authenticity and official status of the skill.- [INDIRECT_PROMPT_INJECTION]: The skill instructions create a vulnerability surface by requiring the agent to analyze external, potentially attacker-controlled data.
- Ingestion points: Build error logs and existing project source code as described in 'SKILL.md' (item 3) and 'references/cli.md'.
- Boundary markers: The instructions do not define any markers or instructions to isolate external data from the system prompt.
- Capability inventory: The skill utilizes shell commands ('ng build', 'ng serve', 'ng test', 'ng add', 'pnpm add', 'ng deploy') which could be misused if a malicious error message or code file is processed.
- Sanitization: No sanitization or validation of the ingested logs or code is specified.- [COMMAND_EXECUTION]: The skill relies on shell command execution for core functionality including building, testing, and serving Angular applications.- [EXTERNAL_DOWNLOADS]: The skill facilitates downloading dependencies and CLI tools via 'pnpm' and 'ng' from official registries. While these are well-known services and trusted organizations, the process involves executing external packages.
Audit Metadata