heating the forge...

@forgecat/msitarzewski_agency-agents_agency-support — Security Risk Report | ForgeCat

msitarzewski_agency-agents_agency-support

Security risk report for @forgecat/msitarzewski_agency-agents_agency-support v0.0.12

Low

Profile references a legitimate public GitHub repository (msitarzewski/agency-agents) with MIT license.
Source attribution is clear and verifiable; no obfuscation or supply-chain spoofing detected.
Commit hash and version information provided for traceability.

Low

Content describes legitimate business agent personas (analytics, finance, legal, support, infrastructure) with standard professional workflows and frameworks.
No instructions to read credentials, exfiltrate data, ignore system prompts, or execute remote payloads.
Code examples and templates are educational/reference material for standard business processes (SQL queries, Python analytics, Terraform IaC, GDPR compliance frameworks) — not malicious guidance or poisoned knowledge.

Details

Findings

Invisible characters detected
agents/support-infrastructure-maintainer.md · SKILL-011
Invisible characters detected
agents/support-legal-compliance-checker.md · SKILL-011

Evidence

Analytics Reporter: 'Develop comprehensive dashboards with real-time business metrics' — standard BI task.
Finance Tracker: 'Implement multiple approval checkpoints for significant financial decisions' — security best practice.
Infrastructure Maintainer: 'Implement comprehensive monitoring before making any infrastructure changes' — reliability best practice.
Legal Compliance Checker: 'Verify regulatory requirements before implementing any business process changes' — compliance best practice.
Support Responder: 'Prioritize customer satisfaction and resolution over internal efficiency metrics' — service standard.

Low

No tool definitions or MCP servers are declared in the profile; permissions are not instantiated.
Agent descriptions outline domain expertise (data analysis, financial planning, legal compliance, infrastructure, customer support) aligned with stated purposes.
Code examples are templates and reference implementations, not executable tool configurations with excessive authority.

Details

Evidence

Profile declares 'Dependencies: None' and no MCP servers or tool hooks.
Each agent persona is a markdown instruction set describing professional workflows, not a tool with shell/file/network permissions.
Infrastructure Maintainer includes Terraform and Prometheus configs as examples, but these are not installed as executable tools — they are documentation.

Low

No MCP servers are defined or referenced in the profile.
No binary execution, network access, or filesystem permissions are configured.
Profile is purely markdown-based agent instructions without server-side integrations.

Details

Evidence

Profile metadata shows 'MCP servers (untrusted data): (none)'.
All content is static markdown and code examples; no dynamic server definitions or tool hooks.