AGENT SKILL

reg-change-triage

Triages a new or changed regulatory instrument into a decision-ready brief covering what changed, date ladder, affected parties, obligations cited to primary source, candidate policies, gaps, and actions. Applicability is proposed with reasoning, never decided. Use when the user says "what does this new rule mean for us", "triage this SR letter", "summarize the impact of this regulatory change", "when do we have to comply", or "which policies does this hit". Do NOT use for enforcement actions, settlements, or consent orders (use enforcement-action-review), one policy against one rule (use policy-gap-analysis), request trackers (use exam-request-response), a vendor SOC 2 (use vendor-risk-review), workbook formatting (use built-in xlsx), or sending email (use built-in outlook-email).

使用场景电子表格
License 许可
MIT
最近更新
2026-11-07
Stars
★ 103
Forks
118

01 前言 Front Matter

name
reg-change-triage
metadata
{ "icon": "DocumentSearch" }
description
Triages a new or changed regulatory instrument into a decision-ready brief covering what changed, date ladder, affected parties, obligations cited to primary source, candidate policies, gaps, and actions. Applicability is proposed with reasoning, never decided. Use when the user says "what does this new rule mean for us", "triage this SR letter", "summarize the impact of this regulatory change", "when do we have to comply", or "which policies does this hit". Do NOT use for enforcement actions, settlements, or consent orders (use enforcement-action-review), one policy against one rule (use policy-gap-analysis), request trackers (use exam-request-response), a vendor SOC 2 (use vendor-risk-review), workbook formatting (use built-in xlsx), or sending email (use built-in outlook-email).

02 资源文件 7

03 SKILL.md 正文

Scope

Do: identify regulatory instruments, build date ladders, extract changed obligations, propose applicability, locate candidate affected policies, build action trackers Do not: state requirements from memory, treat proposed rules as obligations, decide applicability, materiality, or compliance, claim candidate policies provide coverage, register, file, or send anything

When NOT to Use

  • Clause-by-clause coverage of one policy against one rule — use policy-gap-analysis instead.
  • A published enforcement action or consent order — use enforcement-action-review instead.
  • A vendor SOC 2 or third-party controls review — use vendor-risk-review instead.
  • Formatting, charting, or restyling an existing workbook — use xlsx instead.

Regulatory Change Triage Brief

Turn one regulatory instrument into a brief a compliance owner can act on: what changed, by when, who it hits, and what the firm must look at.

You propose. A qualified human decides. Applicability, materiality, and required action are determinations — see house-rules.md § 2.


Read first

Read these in full before step 1. They are binding; this body only summarises them.

  • references/workflow.md — the full text of every step below, with the guardrails, source resolution, and footer detail.
  • references/house-rules.md — destination gate, determination gate, authority tiers, markers, footers, and § 12 data minimisation (every artifact, search, and chat reply, declines included).
  • references/reg-source-access.md — routing, the codification gap, precision filtering, the date ladder.
  • references/document-extraction.md — if the instrument was uploaded.
  • references/triage-outline.md — default brief and tracker structure: the output format.
  • references/self-check.md — the final self-check (step 10).

Workflow

1. Identify the instrument precisely

Detail: workflow.md § 1. Agency · title · document number · type · authority tier, before anything else.

  • A proposed rule is not law. Its PROPOSED - not in force (G1) line goes right after the whole header block (title, Destination, Prepared, Status), never between them.
  • Declines are plain: <category> withheld under data minimisation., with no statement of what SAR confidentiality or any law requires.
  • Never say what does not exist ("has not been published", "is not yet final"). Say what the instrument you hold is; only after a Federal Register search this session write NOT FOUND - Federal Register API has no record of a final rule for <RIN> as of <date> (Gn).
  • Ambiguous instrument: list the candidates and ask; never pick one silently.

2. Retrieve it

Detail: workflow.md § 2. Connector → direct route → browser → ask for an upload → NOT RETRIEVED, per reg-source-access.md § 1. Never answer from memory. Filter searches by agency, type, and docket.

3. Build the date ladder

Detail: workflow.md § 3. Published · Effective · Compliance date · Mandatory compliance date, plus the comment deadline: every row always shown, never collapsed. A row with no stated date leads with NOT FOUND - instrument states no <which> date (Gn), then any quoted relative or proposed wording. Never "n/a"; never infer a compliance date from the effective date.

4. Extract what changed

Detail: workflow.md § 4. CFR part and section, prior text, new text, both quoted verbatim. Run the codification check before calling codified text current; no prior text → GAP — prior text not established. While the definitions section is not retrieved, every entity term in Prior text and New text quotes, preamble quotes included, reads [covered entity].

5. Derive obligations

Detail: workflow.md § 5. One obligation per row: what must be done · who it binds · trigger or threshold · the date it bites · citation · tier.

  • A proposed rule's obligations carry PROPOSED - not in force, one G-bullet naming every row.
  • Guidance is a supervisory expectation, never "required by". Thresholds are quoted exactly.
  • Quotes and paraphrases use [covered entity] while the definitions section is not retrieved.
  • Bites on is the DATES compliance date covering the row, with its DATES citation, or NOT FOUND - instrument states no compliance date for <row> (Gn). Never the effective date.
  • Capture every exemption and exclusion.

6. Propose applicability — never decide it

Detail: workflow.md § 6. For each obligation, propose the business lines, products, entities, or jurisdictions affected, with the reasoning and the text it rests on, marked NOT DETERMINED — compliance owner decides. Retrieve the coverage-defining provision first; if it was not retrieved, the reasoning reads exactly none - defining provision not retrieved. No hedge words. Name any missing fact (for example total assets) with its marker. State each scope ambiguity as an open question with its decider. Never "will require you to".

7. Locate candidate affected policies

Detail: workflow.md § 7. Search M365 for policies touching each obligation: name · version · approval status · approval date · location · why candidate (the topical link only). Nothing found → NOT FOUND - <repository> has no record of a policy for <obligation> (Gn). Each rejected document gets its own GAP - scope mismatch, not a candidate bullet. A candidate is never coverage; hand clause-level work to policy-gap-analysis.

8. Build the action list

Detail: workflow.md § 8. Proposed actions only: the obligation served · a suggested owner with its basis · the driving date · what closes it. No basis → only NOT FOUND - owner not identified (Gn).

9. Emit the tracker and draft handoff

Detail: workflow.md § 9. Build the workbook with xlsx per triage-outline.md. A GRC connector is read for pre-fill only; the deliverable is a draft export, and the register update is the human's action. No connector → NEEDS FEED — GRC change register.

10. Final self-check, then count

  1. Save the artifact text, the chat reply (reply.md), and the search log (searches.md: one REQ | <source> | ok / failed / not attempted line per request and one HIT | <query> | <document> | used / rejected (Gn) / duplicate (Gn) / set aside line per document returned).
  2. Run python scripts/lint_deliverable.py <artifact> --reply reply.md --searches searches.md. Fix every ERROR, resolve every CHECK, and re-run until it prints ERROR: 0 (house-rules.md § 13).
  3. Work through every item in references/self-check.md, including the ones the script cannot judge.
  4. Build the footer counts from the G-list and D-list last, then recount.

Guardrails

Full list: workflow.md, Guardrails.

  • Never state a requirement from memory or cite anything reached only through a snippet.
  • Never present a proposed rule as an obligation or guidance as a legal requirement.
  • Never collapse the four dates; never report codified text as current without the check.
  • Never decide applicability, materiality, or compliance; no hedge words on a determination.
  • Never fabricate a document number, docket, section, date, owner, or figure.
  • Never claim a save succeeded when it failed. Nothing is registered, filed, sent, or committed in a source system.

Output footer

Per house-rules.md § 9 and workflow.md, Output footer: destination · instrument and identifiers · route · tier · coverage · codification check and result · policy repository searched and found · connectors · open items by marker type · determinations outstanding · draft export, no source-system update · files and save result · template used.

Example

Input: FinCEN proposed rule RIN 1506-ABCD; search AML policy owners.
Output row: Obligation 1 - PROPOSED - not in force (G1); comment deadline 2026-11-12.
Candidate policy: AML Program Standard v4; Determination: NOT DETERMINED - compliance owner decides (D1).
Owner: NOT FOUND - owner not identified (G2).
No match: NOT FOUND - Compliance/Policies has no record of a policy for O2 (G3).
D-bullet: D1 - NOT DETERMINED - does O1 reach the firm - compliance owner decides
Handoff: Draft register export prepared; no source-system update performed.

Reference files

  • references/workflow.md — full step text, guardrails, source resolution, footer detail.
  • references/self-check.md — the final self-check.
  • references/house-rules.md — gates, tiers, markers, footers, data minimisation, the lint step.
  • references/reg-source-access.md — routing, codification gap, precision filtering, date ladder.
  • references/document-extraction.md — reading an uploaded instrument or policy.
  • references/triage-outline.md — default brief structure and tracker columns.

04 同仓库更多 Skills

microsoft
microsoft/cat-agent-skills

stakeholder-and-decision-map

Mines customer interactions to identify stakeholders, infer their likely roles and engagement, map the decision process, and highlight who or what may be missing. Use when asked to "map a deal's buying group", identify influential or missing stakeholders, "assess who has gone quiet", or "explain how the customer will decide". Do not use for phased engagement plans, contact sequences, choosing or drafting outreach to the single next contact, seller or team performance reviews, or contacting, scheduling, or changing records.

★ 103
microsoft
microsoft/cat-agent-skills

sales-progress-update

Creates or refreshes a sales progress update for a manager or team about work done, deal movement, pending outcomes, supported forecast position, and help needed. Use when asked to "create my weekly sales update", "refresh my sales progress deck", "report my forecast position", "where do I stand against my quarter forecast", or "catch my manager up on committed deals". Do NOT use for non-sales project updates, a since-date deal change digest, momentum or blocker analysis, meeting prep, or outreach drafting.

★ 103
microsoft
microsoft/cat-agent-skills

submission-assemble

Put the checked CSR or 510(k) sections together in template order with a reviewer-facing open-items register, as a draft package for human sign-off. Use when the user asks "assemble the CSR draft", "put the module together for review", "build the review package with the open items on top", "compile the 510(k) sections in order" or "package this up for the submission lead". Do NOT use for assembling an unchecked draft - use consistency-check instead. Do NOT use for loading protocol, SAP or TLF sources - use source-ingest instead. Do NOT use for compiling general documents or slide decks - use the built-in docx or pptx skill instead. Do NOT use for filing or submitting to a regulator - use the sponsor's validated gateway, operated by a human, instead. Guardrail: draft for human approval only (L1); status is draft_clean or draft_open_items, never submitted.

行政与办公人员文档处理演示文稿
★ 103
microsoft
microsoft/cat-agent-skills

source-ingest

Pull every number, date and population definition out of a study's protocol, SAP, TLF tables, investigator brochure and prior approved text into one cited fact map, keyed per config/field-dictionary.json, before any CSR or 510(k) section is written. Use when the user asks "here's the locked TLF set and SAP, set up the CSR", "load the protocol and tables before we start drafting", "what figures do we have from the database lock?", "pull the disposition counts out of Table 14.1.1" or "ingest the IB for the 510(k)". Do NOT use for drafting or assembling submission sections - use section-draft or submission-assemble instead. Do NOT use for reconciling a patient's medications - use the discharge-care-transition plugin instead. Guardrail: reads and cites sources only; never estimates, rounds or invents a fact.

★ 103
microsoft
microsoft/cat-agent-skills

section-plan

Work out which sections a CSR or 510(k) requires, from the templates in config, check each one against the cited fact map, and name what is missing before a word is drafted. Use when the user asks "what do we still need before we can draft the CSR?", "which sections are ready to start?", "what's blocking the performance data section?", "do we have everything for the synopsis?" or "map our tables to the E3 structure". Do NOT use for writing section language - use section-draft instead. Do NOT use for cross-checking a finished draft - use consistency-check instead. Do NOT use for outlining general Word documents or business reports - use the built-in docx skill instead. Only structures with a template in config are supported. Guardrail: plans only; never marks a section ready on a missing fact and never invents a structure.

文档处理
★ 103
microsoft
microsoft/cat-agent-skills

section-draft

Write or revise the language for one planned CSR or 510(k) section - synopsis, efficacy, safety, discussion, device description - using only figures already in the cited fact map, each with its TLF or SAP citation. Use when the user asks "draft the Synopsis from our locked tables", "write up safety from Table 14.3.1", "give me a first pass at the discussion", "draft the device description for the 510(k)" or "tighten the wording of the Discussion section". Do NOT use for supplying a number not already in source_facts - use source-ingest instead. Do NOT use for verifying a draft's figures - use consistency-check instead. Do NOT use for appeal letters, clinical notes or discharge summaries - use the matching HLS plugin instead. Guardrail: drafts for human approval only (L1); never asserts an unsourced number or harmonises two figures.

★ 103