fundraise-prep

Warn

Audited by Gen Agent Trust Hub on Jun 14, 2026

Risk Level: MEDIUMPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • [PROMPT_INJECTION]: The skill directly interpolates user-provided $ARGUMENTS into the system role definition. This allows an attacker to supply a company name that contains instructions to override the agent's behavior or ignore previous constraints.
  • [DATA_EXFILTRATION]: The skill accesses highly sensitive file paths including data/cfo/cap_table.json and the data/cfo/ directory containing financial metrics and forecasts. Access to equity and financial information presents a risk of sensitive data exposure if the agent is compromised.
  • [PROMPT_INJECTION]: The skill ingests data from various local directories without implementing boundary markers or instructions to ignore embedded commands, exposing it to indirect prompt injection. * Ingestion points: data/cfo/, data/gtm/, data/product/, and data/cfo/fundraise/. * Boundary markers: Absent; no delimiters are used to separate ingested data from agent instructions. * Capability inventory: Reads and processes local files to generate assessment reports. * Sanitization: Absent; the skill does not validate or filter the content of the ingested files.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jun 14, 2026, 10:41 AM
Security Audit — agent-trust-hub — fundraise-prep