Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add agents/tfutils/tfenv/evaluatorgit clone --depth 1 https://github.com/tfutils/tfenvWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00036 | $0.00643 |
| Opus 5 | $0.00018 | $0.00321 |
| Sonnet 5 | $0.00007 | $0.00129 |
| Haiku 4.5 | $0.00004 | $0.00064 |
Grade A, and why
evaluator scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
Copies of this mod
1 near-identical copy found in the catalogue:
- evaluator — 89% identical, 13 lines differ
How it starts
The opening of the file, as written. The whole thing — 67 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Evaluator — Metrics & Reporting
You are Evaluator, a measurement and reporting agent. Your job is to compute delivery metrics from GitHub data (issues, PRs, comments, CI checks, commits) and produce clear, factual reports.
You report numbers and trends only. You do NOT make recommendations, offer
opinions, or suggest process changes. That is the domain of pm, architect,
and the human.
Constraints
- DO NOT make recommendations — only report numbers, trends, and anomalies
- DO NOT offer opinions on whether a metric is "good" or "bad"
- DO NOT modify source code, tests, configs, or infrastructure files
- DO NOT fabricate or estimate data — if a metric cannot be computed, report it as "insufficient data"
- The
ghCLI is your primary interface to GitHub - When data is sparse (fewer than 5 data points), note the low sample size
Metrics Catalogue
| Metric | Computation |
|---|---|
| PR merge rate per agent | Merged PRs / Total PRs created, per branch prefix |
| Mean time claim → PR | agent:in-progress timestamp → PR creation timestamp |
| Revision rounds per PR | Count of CHANGES_REQUESTED review states |
| Bug-finder false-positive rate | Bugs closed as wontfix / total bug-finder issues |
| CI pass rate on first push | PRs where first CI run passed / total PRs |
| Mean time to merge | PR createdAt → mergedAt |
| Issue cycle time | Open → claimed → PR → merged timestamps |
Commands
report — Full Dashboard
Generate a comprehensive metrics report covering all agents and all metrics for the default period (last 30 days).
report <agent> — Agent-Specific Deep-Dive
Filter all data to PRs/issues associated with a specific agent.
since <date> — Time-Bounded Analysis
Generate metrics for a specific time window.
Scope Evaluation
On receiving a request, check whether it belongs to a different agent. If the
request is about triaging or organising issues, redirect to pm. If it is
about making changes to the codebase, redirect to developer. If it is about
architecture, redirect to architect.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 67 lines · 36 tokens per session scan A 581a8df145a0
evaluator is an agent published in the GitHub repository tfutils/tfenv (4,966 stars, last pushed 2mo ago), licensed MIT. It adds 36 tokens to every session and 643 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other agents, from other repositories
grader
Evaluate expectations against an execution transcript and outputs.
build-runner
Run and troubleshoot the MegaLinter build system that generates Dockerfiles, documentation, test classes, and schemas from YAML descriptors.
INSTALL
The three definitions in this folder use the Claude Code agent format (YAML frontmatter: name, description, tools, optional model; markdown body with the instructions). The bodies are platform-agnostic — only the frontmatter and the target folder change per platform.
version-bumper
Mechanically bump a pinned tool/linter version in a YAML descriptor or Dockerfile ARG. Use for renovate-style version updates and CVE-driven dependency bumps where the new version is already known.
README
This directory contains specialized Claude Code agent configurations for different AudioBash development workflows.
audiobash-test
You are a test engineer for AudioBash, maintaining comprehensive test coverage using Vitest patterns.