Tribunal Usage

AdvisoryAudited by Static analysis on Apr 30, 2026.

Overview

No suspicious patterns detected.

Findings (0)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

Tribunal may block or warn on file writes, tests, and agent handoffs after it is initialized.

Why it was flagged

The tool is intended to enforce quality gates, but the documented hooks can affect broad development actions and agent workflow behavior.

Skill content
Once installed (`pip install tribunal && tribunal init`), it hooks into every file write, test run, and agent interaction.
Recommendation

Run Tribunal initialization only in projects where you want these quality gates, and review its configuration before using strict enforcement.

What this means

A plugin pack from an untrusted source could change quality rules or hook behavior in the project.

Why it was flagged

The guide supports installing plugin packs from a registry, direct URL, or local directory; those external packs are not part of the reviewed artifact.

Skill content
tribunal install https://github.com/org/custom-pack   # direct URL
tribunal install ./local-pack/         # local directory
Recommendation

Install Tribunal and plugin packs only from trusted sources, review pack contents, and pin versions where practical.

What this means

Local audit logs may reveal project structure, agent activity, and details about blocked or failed checks.

Why it was flagged

The tool stores persistent audit information that can include file paths, agent IDs, and event details for later review.

Skill content
All hook events logged to `.tribunal/audit.jsonl` ... Fields: `timestamp`, `hook_name`, `file_path`, `outcome`, `duration_ms`, `agent_id`, `detail`
Recommendation

Treat `.tribunal/audit.jsonl` as project-sensitive data, avoid committing it unless intended, and review retention/sharing practices.

What this means

Claude Code sessions connected to the Tribunal MCP server may be able to query audit data or request file checks.

Why it was flagged

The guide describes MCP-accessible tools that can expose audit information and run checks on file paths from Claude Code sessions.

Skill content
Tribunal exposes MCP tools queryable by any Claude Code session: ... `tribunal_audit` ... `tribunal_check_file` — run file_checker on any path
Recommendation

Enable the MCP connector only for trusted workspaces and sessions, and verify what paths and audit data the server exposes.

NoteMedium Confidence
ASI10: Rogue Agents
What this means

After setup, Tribunal may continue recording lifecycle events and enforcing gates during later sessions.

Why it was flagged

The documented hook continues to run at lifecycle events and records session summaries as part of Tribunal's persistent project behavior.

Skill content
`SessionEnd` | `session_end` | Writes session summary to audit log
Recommendation

Confirm Tribunal is enabled only where wanted, and use its status/doctor/configuration commands to understand active hooks.