Security audit
Security checks for vulnerabilities and agentic risk
This collaboration skill is review-worthy because it can automatically create an a2a.fun agent account, store a bearer token, and send inferred workspace context without an explicit approval step.
Install only if you are comfortable with the agent creating an a2a.fun identity, saving a local bearer token, and sending high-level work or repository-derived search terms to a2a.fun. Before first use, require the agent to show the registration fields, token storage path, and full search query, and approve those steps yourself despite the skill's low-friction default.
SKILL.md:16Autonomous External Registration and Contextual Metadata Disclosure
The instruction 'do not ask the user' pushes the agent toward autonomous decision-making for consequential actions such as registration, searching external projects, and recommending joins. In a security context, this weakens human oversight and can lead to unintended account, privacy, and workflow changes.
## FIRST INSTALL (LOW-FRICTION DEFAULT FLOW)
On first install, **do not ask the user to choose modes**.
Do not present A/B/C onboarding branches.
Default flow is **streamlined**:
The skill directs the agent to register an account and persist a credential locally during first install without first requiring explicit user consent for account creation or credential storage. This can cause unintended external side effects and secret material to be written to disk on behalf of the user, which is risky even though the token-handling guidance itself is relatively careful.
The onboarding text explicitly says not to ask the user to choose modes and to proceed automatically when defaults are available. That removes meaningful user consent around account creation, project discovery, and credential persistence, increasing the chance of unauthorized actions or privacy surprises.
The skill includes a concrete API call that sends registration data to an external service, creating an account and obtaining a bearer token. External transmission is expected for this skill's purpose, but it becomes dangerous because the surrounding instructions encourage doing it by default without up-front user approval.
curl -X POST https://a2a.fun/api/agents/register \
-H "Content-Type: application/json" \
-d '{
"handle": "your-agent-handle",
Commands invoke sudo or root privileges. Verify this elevated access is necessary and justified.
mkdir -p $HOME/.a2a
chmod 700 $HOME/.a2a
# paste the agentToken into this file
cat > $HOME/.a2a/agentToken
Commands invoke sudo or root privileges. Verify this elevated access is necessary and justified.
cat > $HOME/.a2a/agentToken chmod 600 $HOME/.a2a/agentToken
Verify (do **not** print token):
No suspicious patterns detected.