# Writing Past Performance > Builds past performance write-ups for a US Federal proposal. Use when the user needs past performance, a past performance volume, past performance citations or references, CPARS or PPQ-based write-ups, or relevance narratives for a federal bid. Selects references for recency, relevancy, and quality; maps each prior contract to the current requirement on the dimensions Section M evaluates; and sub… - **Type:** Skill - **Install:** `agentstack add skill-danielkinneyspears-federal-proposal-skills-writing-past-performance` - **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior - **Seller:** [danielkinneyspears](https://agentstack.voostack.com/s/danielkinneyspears) - **Installs:** 0 - **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills) - **Latest version:** 0.1.0 - **License:** Apache-2.0 - **Upstream author:** [danielkinneyspears](https://github.com/danielkinneyspears) - **Source:** https://github.com/danielkinneyspears/federal-proposal-skills/tree/main/skills/writing-past-performance ## Install ```sh agentstack add skill-danielkinneyspears-federal-proposal-skills-writing-past-performance ``` Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent. ## About # Writing Past Performance Build the past performance write-ups: the evidence that the offeror has done work like this before and done it well. Federal evaluators assess past performance for three things (recency, relevancy, and quality) and they assess relevancy strictly, against the current requirement. A past performance volume that lists impressive contracts without proving they resemble *this* work earns far less than the contracts deserve. This skill writes to the three tests. ## When to use this skill Use this skill to develop the past performance volume or past performance citations for a federal proposal. It works from a library of the offeror's (and teaming partners') prior contracts. It does not write key-personnel experience — that is `developing-key-personnel`. Past performance is about *organizational* performance on prior contracts. ## Inputs **Required:** - The solicitation's past performance instructions (Section L) and evaluation criteria (Section M) — they define recency, relevancy, the reference count and format, and how quality is judged. - A past performance library in `source/` — the prior contracts available to cite, with their facts (customer, value, period, scope) and any performance records (CPARS ratings, PPQs, award fees, metrics). **Preferred upstream artifacts:** - `10-compliance-matrix.md`: the past performance requirement rows. - `05-win-strategy.md`: past performance is often a major proof source for win themes; coordinate. **Without performance records**, the write-ups can still describe relevance, but flag every unsubstantiated quality claim — quality with no evidence is weak. Read `../../shared/glossary.md`, `../../shared/federal-solicitation-primer.md`, and `../../shared/pursuit-workspace.md` if not already read this session. ## Intake Ask these as a numbered list. Skip what the solicitation answers. 1. **The rules.** What does Section L require — how many references, what format, what page or word limit each, prime vs. subcontractor references, the recency window (how far back counts)? 2. **The relevancy definition.** How does Section M define relevancy — on scope, size/dollar value, complexity, contract type, customer, place of performance? What does it count as "similar"? 3. **The library.** What prior contracts are available to cite, and for each: customer, value, period of performance, scope, the offeror's role (prime/sub), and any performance evidence? 4. **Quality evidence.** What performance records exist — CPARS ratings, PPQs, award-fee history, customer commendations, quantified outcomes? 5. **Known problems.** Did any candidate reference have performance problems? If so, what happened and what was the corrective action? ## Workflow ``` Past Performance: - [ ] Step 1: Establish the recency, relevancy, and quality rules - [ ] Step 2: Score and select references - [ ] Step 3: For each reference, build the relevance mapping - [ ] Step 4: Substantiate performance quality - [ ] Step 5: Draft each write-up - [ ] Step 6: Address any performance problems honestly - [ ] Step 7: Write the artifacts and update the matrix ``` **Step 1: The rules.** From Section L and M, pin down exactly what recency, relevancy, and quality mean for *this* solicitation, and the reference count and format. Use `references/past-performance-method.md`. **Step 2: Select references.** Score every library contract on recency, relevancy, and quality against the Step 1 rules. Select the set that maximizes all three, within the allowed count. The strongest reference is the most *relevant* one with strong quality — not the largest or most famous. Prefer references where the offeror was prime unless the solicitation values subcontractor experience equally. **Step 3: Relevance mapping.** For each selected reference, build an explicit mapping: dimension by dimension (scope, size, complexity, contract type, customer, place of performance), show how the prior contract resembles the current requirement. This mapping is the heart of the write-up. Do not make the evaluator infer relevance — state it, on the dimensions Section M uses. **Step 4: Substantiate quality.** Attach the performance evidence: CPARS ratings, PPQ scores, award fees earned, quantified results, commendations. Make quality verifiable, not asserted. **Step 5: Draft.** Draft each write-up using `templates/past-performance-writeup.md`: the contract facts, a concise scope description, the relevance mapping, and the quality evidence — within the format and length Section L allows. **Step 6: Address problems.** If a selected reference had a performance problem, address it honestly: what happened, the corrective action, and the result. Evaluators often have access to CPARS; a problem the offeror omits and the evaluator finds independently is far more damaging than one the offeror explains. If a reference's problems are severe, the better move is usually to select a different reference. **Step 7: Write and update.** Save each write-up in the workspace's `14-past-performance/` directory. Update the past performance rows of `10-compliance-matrix.md`. ## Output One file per reference in `14-past-performance/`, plus updated matrix status. Present the user with: the selected references and why each was chosen, the relevance strength of each, any quality gaps, and any performance problem that needs a decision. ## References - `references/past-performance-method.md`: recency, relevancy, and quality; scoring and selecting references; the relevance mapping; substantiating quality; and handling performance problems. - `templates/past-performance-writeup.md`: the per-reference write-up template. ## Guardrails - **Relevance is the discriminator.** Lead every write-up with how the prior work resembles *this* requirement, on Section M's dimensions. An impressive but non-relevant contract scores low. - **Substantiate quality.** Use CPARS, PPQs, award fees, metrics. Replace "performed excellently" with the rating that proves it. - **Never fabricate.** Do not invent contracts, ratings, values, or outcomes. Past performance is verifiable; fabrication is a disqualifying integrity failure. - **Honesty about problems.** A known problem is explained with its corrective action, or the reference is replaced — never hidden. - **Respect the format.** Reference count, format, and length limits in Section L are hard rules. ## Source & license This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code. - **Author:** [danielkinneyspears](https://github.com/danielkinneyspears) - **Source:** [danielkinneyspears/federal-proposal-skills](https://github.com/danielkinneyspears/federal-proposal-skills) - **License:** Apache-2.0 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-danielkinneyspears-federal-proposal-skills-writing-past-performance - Seller: https://agentstack.voostack.com/s/danielkinneyspears - 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%.