Skip to content
Back to skills

Legal Response

ASecurity

Draft a careful response to a regulator, authority, or counterparty communication — agency inquiry, enforcement letter, formal investigation notice, data-subject access request, or counterparty legal demand. Use for monitor letters, regulator findings, on-the-record requests, DSARs, official complaints, and internal incident escalations. Produces the analysis bundle the orchestrator turns into a draft reply.

  • 10 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added August 31, 2026
ai-agentsgo

Security analysis

A100/100

Scanned August 31, 2026

npx -y skills add WenzhuoXu/lawgent --skill legal-response --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Legal Response?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Legal Response
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/wenzhuoxu-legal-response/badge)](https://www.skillsdirectory.com/skills/wenzhuoxu-legal-response)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: legal-response
description: Draft a careful response to a regulator, authority, or counterparty communication — agency inquiry, enforcement letter, formal investigation notice, data-subject access request, or counterparty legal demand. Use for monitor letters, regulator findings, on-the-record requests, DSARs, official complaints, and internal incident escalations. Produces the analysis bundle the orchestrator turns into a draft reply.
argument-hint: "<incoming communication topic or attachment>"
allowed_connectors: [ecfr_search, federal_register_search, govinfo_search, eurlex_search, "pkulaw_*", flk_npc_search, ethiopia_law_search, ethiopia_law_fetch]
rag_collections: [general]
---

# /legal-response — Authority / Counterparty Response

Draft analysis for responses to a regulator, authority, or counterparty
demand. The skill produces the underlying claims, options, and limits; the
orchestrator authors the final letter or filing.

**Not legal advice.** Drafts must be reviewed by qualified counsel before
service / filing.

## Output Contract (binding)

Same as `/playbook/general_playbook.md` §4 and §5. Emit exactly:

````
## Findings
- [self-contained claim, ONE sentence, ≤60 English words, inline pinpoint `[Source, art./§/p.]`] (label: treaty | regulation | case | guidance | local-law | best-practice | risk | drafting)
- ...

## Out of scope
- ...

## Sources
- [source name] — [URL or local path] — supports findings #X[, online-checked: yes/no, pinpoint: confirmed/unavailable]
````

Hard ceiling **≤ 2000 characters**. No tables, walkthroughs, action
matrices, or sanity-check sections.

## Inputs

Provide the inbound communication or its key facts:
- Sender + authority basis (statutory section, treaty article, contract
  clause).
- Request scope (information, document production, attestation, written
  position).
- Response window + service mode (electronic, certified mail, in-person,
  formal filing).
- Jurisdiction + venue (which body, which body's procedural rules).
- Privilege posture (privileged / discoverable / mixed).

## Workflow

1. **Classify the communication**: regulator inquiry / enforcement / on-the-
   record request / DSAR / counterparty demand / discovery / subpoena.
2. **Identify the controlling authority** under each implicated
   jurisdiction. Confirm the sender's statutory or contractual basis.
3. **Map duties and limits**: what must be answered, what may be answered,
   what must not be disclosed (privilege, state secrets, third-party
   confidentiality, contractual confidentiality), and what carveouts apply
   (regulator disclosure, court order, mandatory reporting).
4. **Surface response options** as separate Findings bullets: full response,
   partial response with objections, request for extension, request for
   clarification, motion to quash / objection.
5. **Reporting / escalation map**: who else needs to know (DPO, regulator,
   insurer, board, outside counsel).
6. **Sanity check** per general playbook §3. No invented citations.

## Reporting / authority taxonomy

Use the right authority bucket per implicated jurisdiction:

- PRC: 监管约谈, 行政询问, 监察建议, 检察建议, 行政复议, DSAR under
  PIPL, 仲裁/诉讼 demand.
- US: agency LOI, formal investigation, civil investigative demand, GAO
  audit letter, subpoena, DSAR/CCPA request.
- EU: regulator notice (national competent authority), GDPR Art. 15 DSAR,
  ESMA / EBA inquiry, court order, DPC investigation.

## Pointers

- Citation + language: `/playbook/general_playbook.md` §1 + §2.
- Sanity check: general playbook §3 + §4.
- Universal regulator disclosure carveout: general playbook §6.
- When the **aviation** pack is active, also apply
  `/domains/aviation/overlays/legal-response.md`.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.