Security risk report for @forgecat/openai_skills_linear v0.0.11
Source Integrity
Low
Profile is sourced from OpenAI's official skills repository (github.com/openai/skills), a trusted upstream.
Metadata clearly attributes authorship and provides commit hash for supply-chain traceability.
Agent Intent
Low
Content describes legitimate Linear workflow management tasks (issue triage, sprint planning, documentation audit) with no instructions to manipulate the AI, exfiltrate data, or bypass security controls.
Setup instructions (MCP configuration, OAuth login) are standard integration steps, not hidden directives or prompt injection attempts.
No guidance poisoning: recommended practices (batch operations, caching, rate-limit awareness) are security-conscious and standard.
Details
Evidence
"Follow these steps in order. Do not skip steps." — procedural guidance for legitimate workflow, not AI manipulation.
"Batch operations for related changes; consider smart templates for recurring issue structures." — standard best practice, not malicious steering.
Permissions
PRM-000Medium
MCP server grants broad issue/project management authority (create, update, list, search) across Linear workspace, teams, and projects — scope is somewhat broader than a single stated function.
No explicit file_write, shell, or file_delete permissions declared, but MCP tool set includes create/update operations that could modify workspace state at scale.
Authority is tied to user's Linear OAuth credentials and workspace permissions, providing some natural boundary.
Details
Evidence
Available tools span issue management, project/team operations, and documentation search — multiple categories of write access.
"For bulk operations, explain the grouping logic before applying changes." — acknowledges potential for large-scale modifications.
MCP Risk
Low
MCP server is a known, official Linear integration (https://mcp.linear.app/mcp) with no hidden instructions or arbitrary binary execution.
Access is scoped to Linear API via OAuth; no unrestricted filesystem or network access.
Tool descriptions are transparent and aligned with stated Linear management purpose.
Details
Evidence
MCP URL is a well-known Linear endpoint; no obfuscated or suspicious host.
Setup requires explicit OAuth login and workspace configuration — no silent or pre-authorized access.
ForgeCat