microsoft/cat-agent-skills is a static website that catalogs reusable instruction sets and related packages for AI agents. People use it to search, filter, rate, and download skills for Cowork, Copilot Studio, and Scout, along with Copilot plugins and Scout automations. The catalogue entries are the skills, instructions, plugins, and settings displayed by the site.
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 microsoft/cat-agent-skills --skill deadline-trackgit clone --depth 1 https://github.com/microsoft/cat-agent-skillsWrote 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/microsoft/cat-agent-skills/deadline-track)<a href="https://agentmods.dev/skills/microsoft/cat-agent-skills/deadline-track"><img src="https://agentmods.dev/badge/skills/microsoft/cat-agent-skills/deadline-track/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/microsoft/cat-agent-skills/deadline-track"><img src="https://agentmods.dev/badge/skills/microsoft/cat-agent-skills/deadline-track.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00186 | $0.01942 |
| Opus 5.5 | $0.00074 | $0.00777 |
| Sonnet 5.5 | $0.00037 | $0.00388 |
| Haiku 4.5 | $0.00019 | $0.00194 |
Grade A, and why
deadline-track 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.
How it starts
The opening of the file, as written. The whole thing — 52 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Deadline Track
Purpose
Guarantee no appeal is missed on a technicality, no denial is quietly dropped, and no coding change is smuggled through.
When to use
Use before appeal-draft (so the letter knows the coding_decision and the clock) and again as the last step of every run, on every denial, whether or not a coding_request exists - typical asks are "when is this appeal due", "how many days are left", "which denial do we work first", "can we recode this and refile at the higher level".
When NOT to Use
Do NOT use when: the ask is to write or reword the appeal letter - use appeal-draft instead; the ask is to find chart evidence - use evidence-assemble instead; the ask is about payer policy - use policy-retrieve instead; the ask is to change, extend or negotiate a payer's appeal window, or to assume one for an unconfigured payer; the ask is to re-order the queue by ease, effort or "quick wins"; or the ask is to apply, file or approve a coding change - this skill decides whether a recode may be honoured, and never performs one.
Inputs
denial.denial_date; denial.payer; denial.recoverable_value; queue[]; config/payer-appeal-windows.json; coding_request (if any); chart_evidence with its citations; policy_citations.
Steps
- Run the engine tool -
python3 base/scripts/deadline_engine.py --question <contract.json> --payers config/payer-appeal-windows.json --out <result.json>- with NO--todayflag, so days are counted from the real current date.--todayexists only to reproduce a demo scenario's snapshot figures; never pass it in a live run, and never compute a deadline by hand or by estimate. - If a coding_request is present, the engine honours it only when chart_evidence.citations names a specific note and date for the requested code (compared on the code itself, so "99215 (higher E/M level)" matches "99215") and supports_requested_code is true; otherwise coding_decision=refused_evidence_gap and the evidence gap is logged.
- The engine ranks queue[] into queue_ranked by deadline ascending then recoverable_value descending, and gives every row a deadline_status.
- The engine flags urgent (days_remaining at or below config.urgent_threshold_days) and expired (days_remaining below zero) for the primary denial AND every queued claim, and echoes the policy clause_ids into deadline_result.policy_basis.
- Payer names are matched to config case-insensitively, so "CONTOSO HEALTH PLAN" uses the Contoso Health Plan window; re-running the engine on a contract that already carries escalations recomputes the engine's own escalations instead of duplicating them, and keeps every other escalation once.
- Report deadline_status first, then the engine's numbers and escalations verbatim to the coordinator. If the engine reports that the contract as_of differs from today, say which date the figures are computed as of.
What ships with it
4 files 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.
- 2d ago First seen · 52 lines · 186 tokens per session scan A b422a89ae48f
deadline-track is a skill published in the GitHub repository microsoft/cat-agent-skills (101 stars, last pushed yesterday), licensed MIT. It adds 186 tokens to every session and 1,942 once invoked, about $0.0007 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 2026-11-06.
deadline-track compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| deadline-trackmicrosoft/cat-agent-skills | Skill | 101 | yesterday | MIT | A | 186 | Claude Code, Codex |
| fto-triageanthropics/claude-for-legal | Skill | 9.9k | yesterday | Apache-2.0 | A | 81 | Claude Code |
| legal-holdanthropics/claude-for-legal | Skill | 9.9k | yesterday | Apache-2.0 | A | 68 | Claude Code |
| specification-writingwanshuiyin/Auto-claude-code-research-in-sleep | Skill | 19k | today | MIT | A | 49 | Claude Code |
| regulatory-research-fallbackHKUDS/OpenSpace | Skill | 7.9k | 2d ago | MIT | A | 20 | Claude Code, Codex |
| terms-of-servicegenspark-ai/genoffice | Skill | 12k | today | Apache-2.0 | A | 32 | Claude Code, Codex |
| x-scorecardx-cmd/x-cmd | Skill | 4.7k | 8d ago | Apache-2.0 | A | 57 | Claude Code, Codex |