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 skills add axiomantic/spellbook --skill auditing-green-miragegit clone --depth 1 https://github.com/axiomantic/spellbookWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/axiomantic/spellbook/auditing-green-mirage)<a href="https://agentmods.dev/skills/axiomantic/spellbook/auditing-green-mirage"><img src="https://agentmods.dev/badge/skills/axiomantic/spellbook/auditing-green-mirage/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/axiomantic/spellbook/auditing-green-mirage"><img src="https://agentmods.dev/badge/skills/axiomantic/spellbook/auditing-green-mirage.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 2 findings, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high Prompt Injection · line 105 Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.Fix: Audit all comments and invisible characters. Remove any instructions that direct the agent to perform unauthorized actions. Use plain, reviewable content.
- high Prompt Injection · line 166 Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.Fix: Audit all comments and invisible characters. Remove any instructions that direct the agent to perform unauthorized actions. Use plain, reviewable content.
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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00107 | $0.03325 |
| Opus 5.5 | $0.00043 | $0.01330 |
| Sonnet 5.5 | $0.00021 | $0.00665 |
| Haiku 4.5 | $0.00011 | $0.00332 |
Grade A, and why
auditing-green-mirage 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 1mo 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.
How it starts
The opening of the file, as written. The whole thing — 259 lines — stays where its author put it; the contents beside it link to each section on GitHub.
This is very important to my career.
MUST:
- Read every test file and verification tool script line by line
- Trace every code path from test/verifier through target code/artifacts and back
- Verify each assertion or check command would catch actual failures (including testing negative controls)
- Identify all gaps where broken code or broken verifier logic would still pass or report clean
- Flag every skipped, xfailed, or conditionally disabled test/check and determine whether the skip hides a real bug
- Audit verification apparatus for tool-blindness (e.g. tools exiting 0 on empty matches, scanner fence/backtick parsing defects, or metrics improving due to broken instruments)
This is NOT optional. Take as long as needed. You'd better be sure.
Dual Scopes
This skill operates across two distinct but complementary scopes:
- Scope 1: Test Suite Audit — Audits test functions, fixtures, assertions, and test suites to detect shallow assertions, over-mocked logic, and tests that cannot fail.
- Scope 2: Verification Apparatus Audit — Audits linters, plan checkers, build-tree verifiers, CI workflows, check scripts, and metric collectors. Detects tool-assertion defects where tools exit 0 on empty matches, scanners develop blind spots, checks pass vacuously, or metrics improve because measuring tools break.
Invariant Principles
- Passage Not Presence - Test/Verifier value = catching failures, not passing. Question: "Would broken code or a missing artifact fail this check?"
- Consumption Validates - Assertions and check commands must USE outputs (parse, compile, execute, assert structure), not just check existence or exit status.
- Complete Over Partial - Full object assertions expose truth; substring/partial checks and unanchored regexes hide bugs.
- Trace Before Judge - Follow test/verifier -> target code/artifact -> return -> assertion path completely before verdict.
- Evidence-Based Findings - Every finding requires exact line, exact fix code, traced failure scenario, or empirical transcript.
- Skipped Tests/Checks Are Silent Failures - A test or check that never runs catches zero bugs. IF skip reason is anything other than a true environmental impossibility (wrong OS, missing hardware), THEN it is unjustified concealment.
- Instrument Output Cannot Certify Instrument - A finding count or green status is an output of a tool, so it can never certify that the tool is functioning. Verify tools using negative controls, mutation tests, or coverage metrics, never by observing clean tool runs alone.
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.
- 1mo ago First seen · 259 lines · 107 tokens per session scan A fa8b4fb04a55
auditing-green-mirage is a skill published in the GitHub repository axiomantic/spellbook (10 stars, last pushed 2d ago), licensed MIT. It adds 107 tokens to every session and 3,325 once invoked, about $0.0004 per session on Opus 5.5. 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 2026-09-03.
Other skills, from other repositories
e2e-testing
Guide for running end-to-end tests of the Qwen Code CLI, including headless mode, MCP server testing, and API traffic inspection. Use this skill whenever you need to verify CLI behavior with real model calls, reproduce user-reported bugs end-to-end, test MCP tool integrations, or inspect raw API request/response…
terminal-capture
Automates terminal UI screenshot testing for CLI commands. Applies when reviewing PRs that affect CLI output, testing slash commands (/about, /context, /auth, /export), generating visual documentation, or when 'terminal screenshot', 'CLI test', 'visual test', or 'terminal-capture' is mentioned.
tmux-real-user-testing
A real-user test workflow for Qwen Code's terminal interface, using tmux to drive the program and capture readable screen snapshots. tmux is a tool for running and observing terminal sessions.
agent-reproduce-align
Use after a Codex or Claude Code feature has been implemented in Qwen Code to run the selected reference agent and Qwen Code under the same scenario, capture HTTP and terminal traces, compare request bodies, tool/function schemas, outputs, and iterate until the reproduced behavior is close enough.
test-failure-explorer
Import and query an existing Vitest 3.2 JSON report to explore test failures, file-level errors and original diagnostic evidence.
verification-before-completion
Use when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always.