Qualification Skill
Write the structured variables for AI prospect qualification. These variables feed any qualification step that scores or filters individual prospects against an ICP.
Output Contract
qualification.md must follow a fixed structure so downstream campaign tooling can parse it reliably. Hard rules this skill must honor:
- Exactly three H2 headings (case-insensitive match):
## Audience Description,## Qualification Criteria,## Disqualification Criteria. - At least 2 of 3 sections must be populated (non-empty). Fewer is a hard parser error.
- No other H2 headings in the file - they'd be silently ignored.
Required Context
- Read
plan.mdfrom the plan folder - Broad ICP definition - Read
search-strategy.mdfrom the plan folder - Search filters already applied. Never restate any of these dimensions in the prompt. - Read
{workspace}/research.md- Client overview, detailed ICP info, competitors ({workspace}= the client campaign workspace, default./cold-email/{slug}/under the current directory)
Core Principles
- Write short, trust the AI - The qualification AI is smart. Give it a brief, plain-language brief and let it reason. A few high-signal lines beat an exhaustive rulebook. No long enumerations of titles, industries, or product names.
- Never repeat the search filters -
search-strategy.mdalready bounds the list (headcount, geography, titles, seniority, industries). Do NOT restate any of it - not in the audience description, not as a criterion. If the filter sets a headcount range, the prompt says nothing about headcount. Qualification only catches what the filters can't see. - Don't over-qualify - Keep it loose. Qualification is a safety net, not a precision filter. When data is incomplete or ambiguous, qualify.
- Disqualification over qualification - Prefer audience_description + a few disqualification_criteria only. Add qualification_criteria only when there's a real positive signal worth confirming.
- Always exclude competitors - The one disqualifier that's always worth including.
See references/qualification-guide.md for detailed guidelines and examples.
Workflow
Step 1: Load Context
- Read plan.md for the broad ICP
- Read search-strategy.md for what's already filtered
- Read client context and research for competitor info and ICP nuances
- Read references/qualification-guide.md
Step 2: Write Qualification Variables
Keep the whole thing short. A sentence or two per section is plenty - the AI fills in the judgment.
Audience Description (always required):
- One plain sentence on who we're targeting. 5-20 words.
- No headcount, geography, title lists, or industries - those live in the filters.
Disqualification Criteria (strongly preferred):
- A few binary deal-breakers the filters can't see (e.g. competitor, wrong business model).
- 2-4 lines. Always include competitor exclusion. Don't pad the list.
Qualification Criteria (optional - usually skip):
- Add only when there's a genuine positive signal worth confirming. Most campaigns don't need it.
Step 3: Write qualification.md
Keep it lean - short lines, no filter dimensions restated:
# Qualification Prompt
## Audience Description
[One plain sentence on who we're targeting. No headcount, geography, titles, or industries.]
## Disqualification Criteria
- [Competitor exclusion]
- [One or two more deal-breakers the filters can't catch, e.g. B2C model, agency/consultancy]
## Qualification Criteria
(Optional - usually skip. Only add a real positive signal worth confirming.)
Output Handling
Workflow mode (plan folder provided):
- Write
qualification.mdto the plan folder root - Confirm: "Qualification prompt saved to {plan_folder}/qualification.md"
Standalone mode:
- Return the output inline to the user
Variable Definitions
The three structured fields used for qualification:
audience_description- short description of who the target audience isdisqualification_criteria- conditions that disqualify a prospect (bullet points)qualification_criteria- optional conditions that qualify a prospect (bullet points)