# Notice Reply Drafter > Drafts a reply to a legal notice received from another party — a demand notice, a cease and desist, a regulatory notice — dealing with each allegation in turn — admitted, denied with the client's own account, denied for insufficient knowledge, or qualified. Use this whenever a user has received a notice and needs to respond — including phrasings like "draft a reply to this demand notice", "respon… - **Type:** Skill - **Install:** `agentstack add skill-rohasnagpal-legal-ai-skills-notice-reply-drafter` - **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior - **Seller:** [rohasnagpal](https://agentstack.voostack.com/s/rohasnagpal) - **Installs:** 0 - **Category:** [Content & Media](https://agentstack.voostack.com/c/content-and-media) - **Latest version:** 0.1.0 - **License:** MIT - **Upstream author:** [rohasnagpal](https://github.com/rohasnagpal) - **Source:** https://github.com/rohasnagpal/legal-ai-skills/tree/main/plugins/advisory/skills/notice-reply-drafter ## Install ```sh agentstack add skill-rohasnagpal-legal-ai-skills-notice-reply-drafter ``` Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent. ## About # Notice Reply Drafter ## What this does Drafts a reply to a legal notice the client has received: addressing every allegation made, stating the client's position on each precisely, and raising any point that affects the whole notice rather than just one allegation. It structures what the client actually says; it does not invent a defence or a fact the client has not given. A response that leaves an allegation unaddressed can, in some contexts, be read as tacit admission, so completeness is treated as a real risk here, not a formality. ## Before you start **The notice being replied to.** The actual text — there is no reply to draft without seeing what was actually alleged. **Which side is instructing.** This is drafted for the recipient of the notice, against the party who sent it. **The client's actual position on each allegation.** What is true, false, or partially true, from the client. This is blocking; the skill structures the client's account, it does not construct a defence from the allegations alone. **Governing law.** Reply timing and formality requirements, and the legal effect of silence or an inadequate reply, can be law-dependent — in some systems, silence in response to a notice itself carries consequence. Extract from any supplied documents or ask; treat compliance with any such requirement as a verification point. Not blocking, ask once and proceed on a reasonable default without it: **tone** — conciliatory, to preserve the relationship, or firm, anticipating that litigation may follow. This shapes register throughout, not the substance of the response. ## Method **1. Read the whole notice once before drafting a single response.** Allegations in a notice routinely interact, and a global point — the claim is time-barred, the notice is defective, the wrong party was named — can dispose of several allegations at once; drafting allegation-by-allegation on a first pass misses this. **2. List every allegation made, numbered, before drafting any response.** Completeness matters on its own terms here: if the client wants to leave an allegation unaddressed, flag the risk of that silence rather than simply omitting a response. **3. State the client's position on each allegation precisely, using only what the client has actually said** — admitted, denied with the client's own account, denied for insufficient knowledge to admit or deny, or qualified. Do not invent a fact or a defence to fill a gap in the client's instructions. **4. Where an allegation is denied, state the client's own version of events specifically, not a bare denial.** A bare denial is weaker and less useful evidentially than a denial coupled with an affirmative account of what the client says actually happened. **5. Raise any global or threshold point prominently**, near the start of the substantive response rather than buried among individual allegations — since a point like limitation or a defective notice can affect the whole document, not just one paragraph of it. **6. Keep factual admissions and legal characterisation separate, even where a fact is conceded.** Admitting "the delivery was three days late" is not the same as admitting "we are in breach," and the reply should never blur the two — concede the fact if it's true, without conceding the legal conclusion the other side is drawing from it. **7. State the client's own position or demand at the end, if any** — rejecting the notice's demand outright, proposing a different resolution, or simply reserving rights pending further information. **8. Flag any reply-timing or formality requirement the governing law may impose** — a mandatory response period, required content, method of service — as a point to verify rather than assume compliance with. ## Output **1. Header.** Recipient of this reply (the party who sent the original notice), sender/client, date, reference to the original notice. **2. Response to each allegation.** Numbered to match the original notice: admitted / denied with the client's own account / denied for insufficient knowledge / qualified. **3. Global or threshold points, if any.** Stated prominently, not folded into the allegation-by-allegation response. **4. Client's own position or demand, if any.** **5. Reservation of rights.** **6. Points requiring verification.** Reply-timing or formality requirements under the governing law, and anything resting on an unconfirmed legal characterisation. ## Do not Do not invent a fact or a defence the client has not actually given. Do not concede a legal conclusion while admitting a fact. Keep the two separate throughout. Do not leave an allegation unaddressed without flagging to the user the risk that silence can carry. Do not draft a bare denial where the client has given their own account of events — use it. Do not assume a reply-timing or formality requirement is satisfied. Flag it as a point to verify. ## Source & license This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code. - **Author:** [rohasnagpal](https://github.com/rohasnagpal) - **Source:** [rohasnagpal/legal-ai-skills](https://github.com/rohasnagpal/legal-ai-skills) - **License:** MIT Install and usage instructions live in the source repository linked above. ## Pricing - **Free** — Free ## Security capabilities Automated source analysis of v0.1.0 — what this tool can access: - **Network access:** no - **Filesystem access:** no - **Shell / process execution:** no - **Environment & secrets:** no - **Dynamic code execution:** no *"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.* ## Versions - **0.1.0** — security scan: passed — Imported from the upstream source. ## Links - Listing page: https://agentstack.voostack.com/l/skill-rohasnagpal-legal-ai-skills-notice-reply-drafter - Seller: https://agentstack.voostack.com/s/rohasnagpal - Browse the marketplace: https://agentstack.voostack.com/browse --- Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.