domain-glossary
Pass
Audited by Gen Agent Trust Hub on Nov 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted conversation history to extract domain terminology and formalize it into a glossary.
- Ingestion points: The skill is instructed to "Scan the conversation for domain-relevant nouns, verbs, and concepts" and, during re-runs, "Incorporate any new terms from subsequent discussion" (found in
SKILL.md). - Boundary markers: There are no specified delimiters or instructions for the agent to ignore potentially malicious embedded commands within the conversation text being analyzed.
- Capability inventory: The skill possesses the capability to read from and write to a local file named
UBIQUITOUS_LANGUAGE.mdwithin the working directory. - Sanitization: The instructions lack any requirement for sanitizing, escaping, or validating the external conversation content before it is interpolated into the generated markdown file. An attacker could potentially embed instructions in the chat to manipulate the output file or attempt to trigger downstream tool actions if the glossary is later processed by other automated systems.
Audit Metadata