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 ui-ai-firstgit 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/ui-ai-first)<a href="https://agentmods.dev/skills/bearded-illirian/trailmark/ui-ai-first"><img src="https://agentmods.dev/badge/skills/bearded-illirian/trailmark/ui-ai-first/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/bearded-illirian/trailmark/ui-ai-first"><img src="https://agentmods.dev/badge/skills/bearded-illirian/trailmark/ui-ai-first.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00145 | $0.02564 |
| Opus 5.5 | $0.00058 | $0.01026 |
| Sonnet 5.5 | $0.00029 | $0.00513 |
| Haiku 4.5 | $0.00015 | $0.00256 |
Grade A, and why
ui-ai-first scanned grade A with 1 finding 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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
are available only via code / curl / SQL and decides per each: How it starts
The opening of the file, as written. The whole thing — 361 lines — stays where its author put it; the contents beside it link to each section on GitHub.
UI/AI-First Protocol
Final protocol of the *-first family. Launched after implementing a large task — before formal closure. Protects against invisible usability debt: features implemented in code but unavailable to a non-developer (everything via curl / SQL / Claude Code).
Input
A completed arch-first task near closure; its task.md, block reports, guides, and code changes.
Output
ui-ai-first-report.md with operations classified A (automate via skill / AI agent) or B (build UI), and a phased roadmap of B items.
Hands off to
ship-first (called from ship-first Step 0.6 when run_ui_ai_first=true).
Introduction
Why it's needed
Every large task leaves "code-only" operations. In reports this doesn't sound like a problem — "feature works", "endpoint in prod". But when an operator wants to use the capability — they hit curl with admin_password in JSON or SQL INSERT into several tables.
3 months later these capabilities effectively exist only on paper — no one but the task author uses them. This is technical debt that looks like "everything's ready".
A/B classification principle
For each "code-only" operation — one of two:
- A — skill / AI agent for rare or technical operations. Quick to build, cheap to run
- B — UI task for frequent operator operations where a skill is insufficient — need a form, button, dashboard
Not every feature needs UI. Not every operation needs a skill.
What it does NOT do
Does not implement findings. Only classifies and forms a prioritized roadmap. Implementation — separate work streams with their own library-first / plan-first.
Step 1 — Get task_path
If passed as argument — use. If not — ask user or propose the last closed task from routing DB.
Verify task is actually closed:
- task.md has blocks with ✅
- ≥1-2
report-NN.mdin task log
If not — stop: "Task is not yet closed. Audit is premature."
Step 2 — Gather list of implemented blocks
Read task.md → all blocks with ✅ or "Done" mark. Ignore planned or cancelled.
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.
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 Changed 14a6478a7ccb
- 1mo ago First seen · 361 lines · 145 tokens per session scan A f9d1a284fa0c
ui-ai-first is a skill published in the GitHub repository bearded-illirian/trailmark (19 stars, last pushed 1mo ago), licensed MIT. It adds 145 tokens to every session and 2,564 once invoked, about $0.0006 per session on Opus 5.5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
stub_distance_skill
a deterministic stub returned by the override engine in unit tests.
generated_skill_1775724185480
Parse XBRL financial documents and extract key taxonomy elements into structured JSON format.
paperclip
Use for Paperclip-managed tasks and heartbeats: reading task context, delivering task documents or files, updating completion or blockers, coordinating or delegating work, and following company governance. Includes control plane API operations for assignments, comments, approvals, and routines.
agentmail
Use your assigned AgentMail inbox to read email tasks, explicitly send or reply, and check delivery. Provided automatically by your inbox assignment.
feishu
Work with Feishu or Lark bots, docs, sheets, bitables, approval flows, and OpenAPI/MCP setup without hardcoding credentials.
interview
Ask one useful structured question at a time only when material product/implementation choices are genuinely missing; remember answers and produce a brief/spec. Discoverable facts should be investigated instead of asked.