Security risk report for @forgecat/anthropics_claude-plugins-official_session-report v0.0.5
Source Integrity
Low
Profile is sourced from Anthropic's official claude-plugins-official repository (https://github.com/anthropics/claude-plugins-official), a trusted first-party source.
Metadata (author, license, commit hash) is transparent and verifiable; no supply-chain obfuscation detected.
Agent Intent
Low
Profile describes a legitimate usage-analytics tool; no instructions to ignore system prompts, exfiltrate credentials, or manipulate the agent's behavior.
Steps are procedural (run analyzer, read JSON, copy template, edit HTML) with no hidden directives or guidance poisoning.
No attempt to weaken security defaults, inject malicious dependencies, or steer future outputs toward attacker goals.
Details
Evidence
Steps 1–5 are straightforward data-collection and report-generation tasks with no obfuscated or adversarial language.
The skill explicitly instructs using absolute paths and local file operations only; no external URL fetches or remote payload execution.
Permissions
PRM-000Medium
The skill instructs the agent to execute a shell command (node ...) and write files to /tmp and the current working directory, which are medium-risk file-write and shell-execution operations.
These operations are broadly consistent with the stated purpose (generating a local HTML report), but shell execution of a bundled, unreviewed script (analyze-sessions.mjs) is broader than strictly necessary.
No alwaysApply=true or wildcard glob rules are declared, and no high-risk unrelated tool categories are requested.
"Edit the output file (use Edit, not Write — preserve the template's JS/CSS)"
MCP Risk
Low
No MCP servers are declared in the profile.
All operations are local (reading ~/.claude/projects, writing to current directory, executing bundled Node.js script); no external host access or unrestricted network calls.
Details
Evidence
MCP servers section is empty; no tool definitions or server configurations present.
ForgeCat