business-dev
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEMETADATA_POISONINGINDIRECT_PROMPT_INJECTION
Full Analysis
- [METADATA_POISONING]: The frontmatter in
SKILL.mdidentifies the author as "pbtc21", which is inconsistent with the expected vendor identity "aibtcdev". This mismatch is documented as a metadata inconsistency but does not impact the operational security of the skill. - [INDIRECT_PROMPT_INJECTION]: The skill handles user-supplied data for CRM management, which presents a surface for indirect injection if that data is sourced from external environments like GitHub or public web pages.
- Ingestion points: Command-line arguments in
business-dev.ts(e.g.,--notes,--pain,--notes,--competition) for theprospect,qualify, andclosesubcommands. - Boundary markers: None. The skill stores input strings directly into a local JSON file without using explicit delimiters to separate user data from instructions.
- Capability inventory: The skill performs file system writes (
writeFileSync) to the~/.aibtc/business-dev/directory and utilizes external capabilities for messaging (x402) and financial transactions (wallet). - Sanitization: The
toIdentifierfunction sanitizes prospect names for use as internal IDs, but other descriptive fields are stored and rendered as raw strings.
Audit Metadata