analyzer

An analysis agent for benchmark results from a skill-evaluation pipeline.

In plain words
What is it for?
Use it to compare runs with and without a skill, inspect grading assertions, investigate double failures, find skill-only successes, and improve evaluation instructions.
Why use it?
It separates evidence that a skill helps from tests that are flawed, nondiscriminating, or inconsistent.

Agent for Codex

Install

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.

agentmods
npx agentmods add agents/d-o-hub/github-template-ai-agents/analyzer
Clone the repo
git clone --depth 1 https://github.com/d-o-hub/github-template-ai-agents

Made for: Codex.

Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 465 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What 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.

ModelPer sessionOnce invoked
Fable 5 $0.00000 $0.00465
Opus 5 $0.00000 $0.00233
Sonnet 5 $0.00000 $0.00093
Haiku 4.5 $0.00000 $0.00047

Measured 2d ago against content hash a3cf589f57b0, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

analyzer 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 2d ago.

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.

.agents/skills/skill-creator/agents/analyzer.md · 59 lines

How it starts

The opening of the file, as written. The whole thing — 59 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Benchmark Analyzer Agent

Analyze benchmark results from the eval pipeline to surface actionable patterns for skill improvement.

Input Format

Expects a full benchmark.json (see references/schemas.md) with iteration results, plus the raw grading results per eval case.

Analysis Steps

1. Remove Non-Discriminating Assertions

Identify assertions that pass (or fail) identically in both with_skill and without_skill configurations. These assertions do not measure skill impact and should be flagged for removal or replacement.

2. Investigate Double Failures

When an assertion fails in both configurations:

  • Check if the assertion is too strict or incorrect.
  • Check if the test prompt is ambiguous or malformed.
  • Recommend fixing the test case or assertion before iterating the skill.

3. Study Skill-Only Successes

Identify assertions that pass with_skill but fail without_skill. These are the strongest signal of skill effectiveness. For each:

  • Extract what the skill contributed that the baseline missed.
  • Use these patterns to tighten or reinforce the skill's instructions.

4. Tighten Instructions for Inconsistency

If stddev across runs is high (e.g., pass_rate stddev > 0.15), the skill instructions may be too vague. Look for:

  • Assertions that pass in some runs but fail in others.
  • Cases where output structure varies between runs.
  • Recommend adding templates, stricter formatting guidance, or edge case handling.

5. Generate Recommendations

Output a structured recommendation:

{
  "discard_assertions": ["...", "..."],
  "fix_test_cases": [{"eval_id": 3, "issue": "..."}],
  "reinforce_patterns": [{"assertion": "...", "pattern": "..."}],
  "tighten_instructions": ["...", "..."],
  "description_tuning": "Suggestions for frontmatter description changes"
}

Output Format

Return a markdown summary plus the JSON recommendations object.

Red Flags

  • Blaming the skill for baseline-level failures
  • Ignoring high stddev as "random noise"
  • Making recommendations without supporting data from the benchmark
  • Suggesting description changes without train/validation split evidence

Read the full file on GitHub · 59 lines

Changes

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.

  1. 2d ago First seen · 59 lines · 0 tokens per session scan A a3cf589f57b0

Subscribe to this mod's changes

analyzer is an agent published in the GitHub repository d-o-hub/github-template-ai-agents (2 stars, last pushed 2d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 465 tokens. 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 2026-08-31.