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 bearded-illirian/trailmark --skill fixture-newgit clone --depth 1 https://github.com/bearded-illirian/trailmarkWrote 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/bearded-illirian/trailmark/fixture-new)<a href="https://agentmods.dev/skills/bearded-illirian/trailmark/fixture-new"><img src="https://agentmods.dev/badge/skills/bearded-illirian/trailmark/fixture-new.svg" alt="Measured on agentmods" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00122 | $0.01549 |
| Opus 5 | $0.00061 | $0.00775 |
| Sonnet 5 | $0.00024 | $0.00310 |
| Haiku 4.5 | $0.00012 | $0.00155 |
Grade A, and why
fixture-new 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 4d 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 — 193 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Fixture-New Protocol
Writes one fixture: fixtures/{skill}/{case}/ holding input.md and
expect.yml.
A fixture answers "did this runtime produce the same shape of work as the reference one" — never "did it produce the same words". Two agents word an artifact differently, and so does one agent twice.
When to invoke
- A skill has no fixture and you want the parity suite to reach it
- An existing fixture covers the happy path and a second case is worth freezing
- Before claiming a runtime is supported for a skill nothing checks
Input
The name of a shipped skill, a short case name, and knowledge of what that skill's artifact looks like.
Output
Two files under fixtures/{skill}/{case}/, plus a demonstration that the new
fixture fails when handed a directory with nothing in it.
Hands off to
$parity-check — the skill that actually runs a fixture end to end.
Step 0 — Locate the workspace root
test -d fixtures && test -d bin && echo "root ok"
Not there → walk up until fixtures/ and bin/ sit side by side.
Step 1 — Name the fixture
Ask for the skill and a case name. Then check both:
ls aihub/.agents/skills/{skill}/SKILL.md # the skill must exist
ls -d fixtures/{skill}/{case} 2>/dev/null # the case must not
Skill missing → stop. A fixture for a skill that does not ship tests nothing and will fail forever, which trains people to ignore red.
Case exists → stop and say so. Overwriting someone's fixture silently is how a suite quietly loses coverage.
Case names read as NNN-short-slug: 001-add-endpoint, 002-empty-input. The
number orders them; the slug says what is being frozen.
Step 2 — Decide the expected shape
Read the skill's own protocol to see what it writes, then fill these:
| Key | What to put | How to find it |
|---|---|---|
artifact |
filename glob, e.g. plan-first-*.md |
the skill's artifact-writing step |
artifact_type |
value in task_artifacts.artifact_type |
the skill's INSERT statement |
sections |
headings that must be present | the skill's artifact template |
gate |
substring proving an approval gate was offered | only if the gate lands in the artifact |
min_table_rows |
minimum rows in the first table | only when the protocol states a count |
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.
- 4d ago First seen · 193 lines · 122 tokens per session scan A 573cf78d6f5f
fixture-new is a skill published in the GitHub repository bearded-illirian/trailmark (19 stars, last pushed 6d ago), licensed MIT. It adds 122 tokens to every session and 1,549 once invoked, about $0.0006 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 2026-09-04.
Other skills, from other repositories
qa
Run scalable, isolated live QA for nac development. The top-level local orchestrator must parse n (default 4), dispatch one setup worker with this skill, copy its n assignment contracts verbatim into exactly n parallel test workers with this skill, then dispatch one aggregate worker with this skill using all test…
chrome-automation
A browser-automation toolkit for controlling Chrome to visit pages, interact with forms, collect data, run tests, and create screenshots or PDFs.
agent-tester
Test agent: dry-run, unit, integration, compatibility.
browser-check
Drive a real browser and come back with a MEASUREMENT rather than an impression - console errors and >=400 responses as counts, computed styles as JSON when appearance is disputed. Use when work touches UI, when a review must verify one, or when fidelity-gate asks for its measurement.
dos-witness-claim
Route subagent claims through independent read-back before another agent relies on them. Use at parallel, pipeline, or synthesis barriers where shipped phases, files, rows, messages, or other effects must be witnessed.
dos-self-improve
Run the DOS self-improvement loop: propose a candidate, verify it in an isolated worktree, measure suite/truth/metric gates, and keep only confirmed gains. Use for recursive improvement; use dos-enforce-tune for policy knobs.