skillscope-e2e

A complete workflow for testing and improving one coding-agent skill on one SkillsBench task. It runs the original skill, analyzes its behavior, rewrites it once when justified, and compares the results.

In plain words
What is it for?
Use it to run the original and optimized task, inspect verifier and analysis artifacts, create one revised SKILL.md when needed, and compare both runs.
Why use it?
It shows whether a skill actually caused a problem and whether a revised version improves adherence without reducing task correctness.

Skill for Claude CodeCodex

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 skills/mathieu0905/skilllens/skillscope-e2e
Any agent
npx skills add mathieu0905/skilllens --skill skillscope-e2e
Clone the repo
git clone --depth 1 https://github.com/mathieu0905/skilllens

Made for: Claude Code, Codex.

Per session 72 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,976 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 $0.00072 $0.02976
Opus 5 $0.00036 $0.01488
Sonnet 5 $0.00014 $0.00595
Haiku 4.5 $0.00007 $0.00298

Measured yesterday against content hash fbf0174b57db, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

skillscope-e2e 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.

skills/skillscope-e2e/SKILL.md · 334 lines

How it starts

The opening of the file, as written. The whole thing — 334 lines — stays where its author put it; the contents beside it link to each section on GitHub.

SkillScope E2E

Use this skill for one selected (taskId, skillRelPath) instance.

The loop is:

select one task + one skill
  -> run original skill on that task
  -> collect trajectory and native verifier result
  -> run SkillScope graph-guided agent judge on the selected skill
  -> gate: pass-through, optimize, or mark not skill-caused
  -> generate exactly one optimized SKILL.md when gated for optimization
  -> rerun the same task with only that selected skill replaced
  -> judge the optimized trajectory and optimized skill
  -> compare native correctness and SkillScope non-compliance

Experiment Unit

Define the instance before running anything:

  • taskId: one SkillsBench task.
  • skillRelPath: one environment/skills/<skill-name>/SKILL.md.
  • trial: default 1 original trial and 1 optimized trial unless the user asks otherwise.
  • slug: one semantic experiment name for this instance.

Prefer one-skill SkillsBench tasks for the first clean loop. A one-skill task is one whose environment/skills directory contains exactly one SKILL.md.

If a task contains multiple skills:

  • Analyze and optimize only the selected skillRelPath.
  • Treat sibling skills as unchanged task context.
  • Do not merge sibling skills into the optimization target.
  • Do not optimize sibling skills in the same run.
  • If the current CLI cannot restrict rerun or propose steps to the selected skill, stop and either choose a one-skill task or add target-skill filtering before continuing.

Versioning Rule

Use git commits for code, workflow, and skill versions. Do not create v1, v2, v8, or similar artifact directories as a substitute for version control.

Use semantic experiment names for artifacts, for example:

  • .skilllens/experiments/skillsbench-codex-gpt55/azure-bgp-single-original
  • .skilllens/experiments/skillsbench-codex-gpt55/azure-bgp-single-optimized
  • .skilllens/experiments/skillsbench-codex-gpt55/right-shift-contract-smoke

Each final comparison must name the git commit, run plan path, original run root, selected skill path, optimized skill root, optimized run root, and both analysis reports.

Read the full file on GitHub · 334 lines

Files

What ships with it

1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

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. yesterday First seen · 334 lines · 72 tokens per session scan A fbf0174b57db

Subscribe to this mod's changes

skillscope-e2e is a skill published in the GitHub repository mathieu0905/skilllens (2 stars, last pushed 2mo ago), licensed MIT. It adds 72 tokens to every session and 2,976 once invoked, about $0.0004 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 2026-08-31.

Related

Other skills, from other repositories

systematic-debugging

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.

obra/superpowers · 21 tokens

next-cache-components-adoption

Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…

vercel/next.js · 95 tokens

babysit-pr

Babysit a GitHub pull request after creation by continuously polling review comments, CI checks/workflow runs, and mergeability state until the PR is merged/closed or user help is required. Diagnose failures, retry likely flaky failures up to 3 times, auto-fix/push branch-related issues when appropriate, and keep…

openai/codex · 114 tokens

imagegen

Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output…

openai/codex · 113 tokens

cpu-profile-analysis

Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…

microsoft/vscode · 71 tokens

next-cache-components-optimizer

Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…

vercel/next.js · 170 tokens