Skip to content
Get packs

Guide · Agent skill

How to use the RFP Responder agent skill (rfp-responder)

Updated · written with AI from the project's own documentation and checked against our catalogue data

RFP Responder is an agent skill: a folder of instructions and Python scripts that an AI coding agent such as Claude can load and follow when a buyer sends a request for proposal, RFI, RFQ, security questionnaire or vendor questionnaire. It is written for bid managers, proposal leads, sales directors and sales engineers who need a structured response strategy rather than finished prose. By the end of this guide you will know how to install the skill, trigger it with a plain request, run its three scripts and read the output: a requirement list, a proof-point matrix with GAPs flagged, win-themes and a bid, partner-bid or no-bid suggestion.

RFP Responderalirezarezvani/claude-skills · MIT

Quick start

  1. 1Download and unzip the skill. Download the ZIP from this skill's page and unzip it into ~/.claude/skills/ for personal use, or into .claude/skills/ in a project folder.
  2. 2Or upload it in the Claude apps. In the Claude apps, open Settings, go to Capabilities, then Skills, and upload the ZIP.
  3. 3Install Python 3. The scripts need Python 3 and use only its standard library, so no extra packages are required.
  4. 4Fill in the intake template. Complete assets/rfp_intake_template.md with deal context and your proof-point library, then save it as rfp_intake.json.
  5. 5Parse the RFP. Run scripts/rfp_parser.py on the RFP markdown or text file to get requirements tagged MANDATORY, WEIGHTED or NICE-TO-HAVE as JSON.
  6. 6Build the proof-point matrix. Run scripts/response_drafter.py to get STRONG, PARTIAL and GAP ratings per requirement, win-theme threading and the GAP audit.
  7. 7Estimate winrate and decide. Run scripts/winrate_predictor.py with your deal context, then take the verdict and GAP audit into the go or no-go review.

What the rfp-responder skill makes the agent do

The skill walks the agent through five questions at the moment a buyer's request lands: what the document is asking, what your true fit is, which win-themes to lead with, what a realistic winrate looks like, and whether to bid.

Each requirement is tagged MANDATORY, WEIGHTED or NICE-TO-HAVE based on cue words in the buyer's text. Each one is then matched to a proof point and labelled STRONG, PARTIAL or GAP. The skill is built to surface gaps openly rather than write around them, so leadership can decide whether to close the gap, partner or walk away.

  • Parses multi-section buyer requirements and scoring criteria if the buyer disclosed them
  • Builds a proof-point matrix mapping each requirement to a verifiable source such as a case study, certification, customer quote, technical attestation or benchmark
  • Proposes 3-5 win-themes that span several requirements
  • Produces a directional winrate estimate with a confidence band and a BID, PARTNER-BID or NO-BID verdict

Install the rfp-responder skill

There are two ways to add the skill. For a personal setup, download the ZIP from the skill's page on this directory and unzip it into your user skills folder, ~/.claude/skills/. To make the skill available only inside one project, unzip it into that project's .claude/skills/ folder instead.

If you use the Claude apps rather than a local agent, upload the ZIP under Settings, then Capabilities, then Skills.

  • Personal install: unzip into ~/.claude/skills/
  • Project install: unzip into .claude/skills/ in your project folder
  • Claude apps: Settings → Capabilities → Skills, then upload the ZIP

Prepare your inputs before you run anything

The skill needs three things: the RFP as a markdown or text file, a proof-point library, and the intake details filled in from the provided template, assets/rfp_intake_template.md. Save the completed intake as rfp_intake.json. The template's JSON skeleton is the canonical input format.

Install Python 3 on the machine where the scripts will run. The scripts use only the Python standard library, so there are no packages to add. Each script accepts --help and --sample, which is the safest way to see what input it expects.

  • RFP document in markdown or plain text
  • Proof-point library: each proof tagged with its type, verifiable source and the requirements it covers
  • Intake file filled in from assets/rfp_intake_template.md, saved as rfp_intake.json
  • Python 3

Trigger the skill with natural requests

You do not need to name scripts or flags. Describe the situation in plain words when the request arrives, and the agent should recognise the trigger. The skill's own description lists the moments it fits: an RFP, RFI, RFQ, security questionnaire, vendor questionnaire or proposal request that needs a structured response.

Good requests name the document and the decision you need. For example: "A buyer sent this 40-page RFP with a 10-day deadline. Parse it, tag the requirements and tell me where we have gaps." Or: "Map our proof-point library to these security questionnaire items and flag anything we cannot back up."

  • "Parse this RFP and tag each requirement as mandatory, weighted or nice-to-have."
  • "Build a proof-point matrix for this RFP and list every GAP."
  • "Suggest three to five win-themes for this bid."
  • "Estimate our winrate and recommend bid, partner-bid or no-bid."
  • "Triage this vendor questionnaire into structured answers."

Run the three scripts step by step

The workflow runs in order. First, the parser extracts requirements, deadlines and submission format constraints into JSON. Second, the drafter combines the parsed RFP with your intake to produce the proof-point matrix and GAP audit. Third, the winrate predictor takes deal context and returns the estimate, factor breakdown and verdict.

The winrate predictor accepts a --profile flag to tune base rates to the kind of buyer. The source names enterprise-software in its example; government, enterprise SaaS and healthcare profiles are described as tuning base rates in the skill's assumptions.

  • Parse: python scripts/rfp_parser.py --input rfp.md --output json > parsed.json
  • Match: python scripts/response_drafter.py --input draft_input.json --output markdown > matrix.md
  • Estimate: python scripts/winrate_predictor.py --input deal_context.json --profile enterprise-software --output markdown
python scripts/rfp_parser.py --input rfp.md --output json > parsed.json python scripts/response_drafter.py --input draft_input.json --output markdown > matrix.md python scripts/winrate_predictor.py --input deal_context.json --profile enterprise-software --output markdown

Read the output and decide

Treat the winrate figure as a directional aid. The skill's own assumptions say the confidence band is always wider than the point estimate suggests, so it should not be the only basis for a pursuit decision. Any estimate below 20% triggers an automatic no-bid recommendation.

The skill does not commit pursuit budget. It hands the parsed RFP, proof-point matrix, GAP audit and winrate estimate to the go or no-go review, where leadership decides.

  • Check every GAP line before anyone drafts a response that touches it
  • Look for win-themes that appear in fewer than two requirements; the drafter flags these as decorative
  • Treat an estimate under 20% as an automatic no-bid recommendation
  • Record who made the final call at the go or no-go review

Safety notes: scripts, files and network

The skill runs local Python scripts. The listing states that it runs scripts and does not use the network. All three scripts are described as standard-library only, so they need no third-party packages. Read the scripts before running them on confidential buyer material, since the RFP and your proof-point library will be processed on your machine.

The skill's rule is that proof points must be verifiable. Do not paste a claim into the library unless you can point to the case study, certification, reference or attestation that supports it.

  • Runs scripts locally; uses no network according to the listing
  • Standard library only: argparse, json, sys, pathlib, re, collections, statistics
  • Keep confidential RFP content on machines you control
  • Unsourced claims become GAPs by design

What people use it for

  • A bid manager triaging a 30-plus page RFP with a short deadline to decide whether it is worth pursuing
  • A sales engineer mapping an existing proof-point library to a buyer's requirement list and finding the gaps
  • A proposal lead triaging a security questionnaire into structured answers with sources attached
  • A director of sales preparing a bid, no-bid or partner-bid review with a documented winrate factor breakdown

Frequently asked questions

Is the rfp-responder skill free?

The skill is published under the MIT licence, as stated on its listing. Check the licence file in the download for the full terms before using it in a commercial setting.

What do I need to run the rfp-responder skill?

You need Python 3, since the scripts use only the standard library. You also need the RFP as a markdown or text file, a proof-point library, and the intake details filled in from the provided template.

How do I install rfp-responder in Claude?

Download the ZIP from the skill's page and unzip it into ~/.claude/skills/ for personal use or into .claude/skills/ in a project. In the Claude apps, upload the ZIP under Settings, then Capabilities, then Skills.

Does rfp-responder write the full proposal for me?

No. The skill states it is not for free-form proposal narrative writing, contract redlines or marketing collateral. It focuses on parsing requirements, mapping proof points, suggesting win-themes and estimating winrate.

How accurate is the winrate estimate?

The skill describes the winrate as a directional heuristic, not a forecast, with a confidence band wider than the point estimate. It should not be the sole basis for a pursuit decision.

What should I do if the skill reports a GAP?

Do not write a claim to cover it. The skill surfaces GAPs for leadership to decide whether to close the gap, partner-bid or no-bid. Inventing a proof point to fill a gap is the failure mode the skill is built to prevent.

New agent skills, repos and AI tools — every week

The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.

Our verdict

Use rfp-responder if you are a bid manager, proposal lead or sales engineer who needs a disciplined bid, partner-bid or no-bid check and an honest list of gaps before a response is written. Skip it if you mainly need free-form proposal writing, contract redlines or marketing collateral, which the skill explicitly excludes. Its winrate figure is a directional aid, so pair it with your own judgement at the go or no-go review.

Download the skill