sales-cycle
Skill0 starsChanged 6 months agoMITView on GitHub
Drive a Sales Cycle end-to-end (Qualify → Scope → Estimate → Propose → Order → Delivery handoff) by gathering deal context from the user up front, then dispatching the right phase agent for each stage, persisting artifacts under sales/<deal-slug>/, and gating between phases with the user. Use when a service provider wants to manage a client opportunity, when someone says "we have a new lead", "we need to write a proposal", "we got an RFP", "help me scope this project for a client", "we need an SOW", or "let's run a sales cycle". Skip when the user already has a signed contract — go directly to /discovery:start or /spec:start instead.
In this repository45 other files it gives its agents.
- AGENTS.mdAGENTS.md
- plugin-v2/AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- plugin-v2/CLAUDE.mdCLAUDE.md
- .github/copilot-instructions.mdCopilot instructions
- .cursor/rules/agents.mdcCursor rule
- agentic-security-reviewSkill
- arc42-baselineSkill
- caveman-commitSkill
- caveman-helpSkill
- caveman-reviewSkill
- cavemanSkill
- create-goalSkill
- design-twiceSkill
- discovery-sprintSkill
- domain-contextSkill
- github-project-setupSkill
- grillSkill
- improve-codebase-architectureSkill
- issue-breakdownSkill
- issue-draftSkill
- issue-pr-syncSkill
- new-adrSkill
- new-glossary-entrySkill
- orchestrateSkill
- portfolio-trackSkill
- product-pageSkill
- project-reviewSkill
- project-runSkill
- project-scaffoldingSkill
- quality-assuranceSkill
- quality-metricsSkill
- record-decisionSkill
- review-fixSkill
- roadmap-managementSkill
- specorator-designSkill
- specorator-improvementSkill
- specorator-onboardSkill
- stock-takingSkill
- tackle-issueSkill
- tdd-cycleSkill
- token-budget-reviewSkill
- tracer-bulletSkill
- ubiquitous-languageSkill
- verifySkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 2 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 41
No active deals → skip to Step 2.
- Distinctive: line 54
Don't ask "should I proceed?" — proceed once answers in.
- Very specific: line 64
Pre-flight: read sales/<slug>/deal-state.md, confirm every upstream artifact complete or skipped.
- Very specific: line 90
Read order.md's delivery_workflow field: discovery or spec.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Sales Cycle
Conductor of Sales Cycle Track in docs/sales-cycle.md. Job: sequence phases + gate between. Never do phase work. Each phase runs in specialist subagent (.claude/agents/); you persist state, surface choices, dispatch.
Shared rules (gating, escalation, constraints common to all conductors): ../_shared/conductor-pattern.md.
Read first
docs/sales-cycle.md— 5-phase sales cycle definition.memory/constitution.md— principles every phase obey (especially Article VII: humans own intent, priorities, acceptance).docs/sink.md— where artifacts live.sales/directory — active deals to resume.
Workflow you drive
| # | Phase | Subagent | Slash command | Artifact |
| — | Bootstrap | — | /sales:start <deal-slug> | deal-state.md |
| 1 | Qualify | sales-qualifier | /sales:qualify | qualification.md |
| 2 | Scope | scoping-facilitator | /sales:scope | scope.md |
| 3 | Estimate | estimator | /sales:estimate | estimation.md |
| 4 | Propose | proposal-writer | /sales:propose | proposal.md |
| 5 | Order | proposal-writer | /sales:order | order.md |
| → | Delivery | orchestrate or discovery-sprint | /spec:start or /discovery:start | downstream workflow |
Steps
Step 1 — Detect resume vs. fresh start
List sales/*/deal-state.md files with status active or on-hold. For each, show: deal-slug | client | current_phase | last_updated. Batch one AskUserQuestion asking user to pick:
- Resume listed deal (recommended-first by
last_updated). - Start new deal.
No active deals → skip to Step 2.
Step 2 — Clarify deal context (single AskUserQuestion, ≤ 4 questions)
New deal, batch one call:
- Deal slug — kebab-case, ≤ 6 words, format
<client>-<project>. Derive from$ARGUMENTSif description given; offer as recommended. - Client name — who is prospect?
- Lead source — inbound / referral / outbound / RFP.
- What we know so far — materials available (RFP document, email thread, meeting notes, problem description)?
Resuming deal, ask: Continue from <next phase> (Recommended) / Re-run <current phase> / Skip to specific phase.
Don't ask "should I proceed?" — proceed once answers in.
Step 3 — Bootstrap (fresh start only)
Invoke /sales:start <deal-slug> <client-name>. Creates sales/<deal-slug>/ + deal-state.md. Don't edit those yourself.
Step 4 — Run phases sequentially
Each phase:
- Pre-flight: read
sales/<slug>/deal-state.md, confirm every upstream artifactcompleteorskipped.complete/skippedpassable;pending/in-progress/blocked= return-to-that-phase signal. - Dispatch slash command for phase (e.g.,
/sales:qualify). Hand off fully — don't duplicate agent's work. - Wait for phase complete + artifact exist.
- Gate with user via single
AskUserQuestion:Continue to <next phase>(Recommended)Pause here— setstatus: on-hold, exit; resume re-invoking/sales-cycle.Re-run <this phase> with feedback— pass free-text as additional context.Mark as no-go— setstatus: closed,current_phase: no-go, record reason indeal-state.md.
Special gates:
- After Phase 1 (Qualify):
pursueverdict needs explicit human confirmation. Present win probability score, top 3 supporting signals, top 3 red flags. Ask:Confirm pursue/Override to no-go/Request more information. - After Phase 3 (Estimate): Present cost range, confidence level, pricing model recommendation. Ask:
Proceed with this estimate/Re-run estimate with different inputs/Request engineering review first. - After Phase 4 (Propose): Present internal review checklist result. Remind user: you must send proposal — not automated. Ask:
Mark as sent/Make revisions first. - After Phase 5 (Order): Present Project Kickoff Brief summary + downstream workflow recommendation. Ask:
Open delivery workflow now/Wait — I'll start delivery separately.
- After Phase 1 (Qualify):
Step 5 — Delivery handoff
User confirms order accepted, wants delivery workflow open:
- Read
order.md'sdelivery_workflowfield:discoveryorspec. discovery: invoke/discovery:start <delivery_slug>. Pass Project Kickoff Brief as input to frame phase.spec: invoke/spec:start <delivery_slug>. Analyst readsorder.mdas mandatory input to Stage 1.
Report full path to new delivery folder, confirm handoff complete.
Step 6 — No-go handling
Any phase, verdict no-go:
- Record no-go reason in
deal-state.mdunderHand-off notes. - Set
status: closed,current_phase: no-go. - Report no-go verdict, reason, any learning informing future qualification/scoping.
No-go = successful outcome — prevented wasted engineering effort. Don't frame as failure.
Constraints (sales-cycle-specific)
Generic conductor constraints + escalation pattern: ../_shared/conductor-pattern.md. Specifics for this skill:
- Never send a document to external parties. Sending is always a human action.
- Never advance Qualify → Scope without a human-confirmed
pursueverdict. - Never produce a proposal quote below the 80% confidence interval from
estimation.md.
References
docs/sales-cycle.md— full methodology: methods, quality gates, deal state machine.docs/sink.md— artifact location map (includingsales/tree).docs/discovery-track.md— downstream track for exploratory mandates.docs/specorator.md— downstream track for defined mandates.memory/constitution.md— Article VII: humans own intent, priorities, acceptance.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

