agentleFS
CodeAlive-AICodeAlive-AI/ceo-ai-os/skills/pipeline-manager/SKILL.md

pipeline-manager

Skill9 starsChanged 6 months agoMITView on GitHub

Manage the sales pipeline: add/update prospects, score with FIT+INTENT+TIMING, track stage transitions, generate status reports. Central orchestrator for all pipeline skills. For signal detection, see signal-scanner. For outreach execution, see outreach-sender.

The file

We split this skill into the rules it gives an agent

9 rules. About 0 are common, 3 are distinctive, and 6 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 6

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Before helping, read MEMORY.md for: current wedge, ICP, competitors, PMF stage.

    line 9
  2. Very specific:

    All prospect data lives in memory/pipeline.json.

    line 20
  3. Distinctive:

    When a new prospect is identified (via signal-scanner, manual, or inbound):

    line 61
  4. Very specific:

    Check if prospect already exists in pipeline.json (match by domain)

    line 63
  5. Very specific:

    If new: create entry with initial data, set stage = signal_detected

    line 64
  6. Distinctive:

    25-49: Add to nurture list

    line 69
  7. Distinctive:

    When new information arrives (enrichment, reply, meeting):

    line 75
  8. Very specific:

    Read current prospect state from pipeline.json

    line 77
  9. Very specific:

    Write back to pipeline.json

    line 82

≈ “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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

[YOUR_PRODUCT] Context

Before helping, read MEMORY.md for: current wedge, ICP, competitors, PMF stage. Pipeline state lives in memory/pipeline.json. Read it at the start of every pipeline operation. Output in Russian (company names, emails, scores in original). For chat platforms: compact pipeline card, max 15 lines.

Pipeline Manager

Central orchestrator for the full-cycle sales pipeline. Manages prospect state, scoring, stage transitions, and reporting.

Pipeline State

All prospect data lives in memory/pipeline.json. Structure:

{
  "prospects": [{
    "id": "company-slug",
    "company": "Company Name",
    "domain": "company.com",
    "contact": {
      "name": "Full Name",
      "title": "VP Engineering",
      "email": "name@company.com",
      "linkedin": "linkedin.com/in/username"
    },
    "scores": {
      "fit": 0,
      "intent": 0,
      "timing": 0,
      "total": 0
    },
    "stage": "signal_detected",
    "signals": [],
    "outreach": [],
    "next_action": { "type": "", "date": "", "detail": "" },
    "notes": "",
    "created": "ISO",
    "updated": "ISO"
  }],
  "stats": {
    "total": 0,
    "by_stage": {},
    "last_scan": "",
    "last_review": ""
  }
}

Operations

1. Add Prospect

When a new prospect is identified (via signal-scanner, manual, or inbound):

  1. Check if prospect already exists in pipeline.json (match by domain)
  2. If new: create entry with initial data, set stage = signal_detected
  3. Score using FIT+INTENT+TIMING (see Scoring below)
  4. Route based on score:
    • 75+: Notify CEO immediately via Telegram. Tag as priority.
    • 50-74: Queue for agent-drafted outreach
    • 25-49: Add to nurture list
    • <25: Monitor, re-score next month
  5. Log via structured-log

2. Update Prospect

When new information arrives (enrichment, reply, meeting):

  1. Read current prospect state from pipeline.json
  2. Update relevant fields
  3. Re-score if new signals detected
  4. Advance stage if transition criteria met
  5. Set next_action with date and detail
  6. Write back to pipeline.json

3. Stage Transitions

signal_detected → researched → qualified → outreach_sent →
follow_up_1 → follow_up_2 → follow_up_3 →
meeting_booked → discovery_done → proposal_sent →
closed_won | closed_lost | nurture

Transition triggers:

  • signal_detected → researched: company-intel + person-lookup + lead-enrichment completed
  • researched → qualified: FIT+INTENT+TIMING score calculated, ≥25
  • qualified → outreach_sent: CEO approved draft, email sent via outreach-sender
  • outreach_sent → follow_up_N: Automatic by cadence (Day 3/5/7/10)
  • Any → meeting_booked: Reply with meeting intent received
  • meeting_booked → discovery_done: After call + discovery-debrief completed
  • discovery_done → proposal_sent: Proposal or demo delivered
  • Terminal: closed_won, closed_lost, nurture

4. Pipeline Report

Generate a pipeline status report. Use for weekly reviews or on-demand.

Report contents:

  • Total prospects by stage (funnel visualization)
  • Top 5 highest-scored prospects with next actions
  • Stale prospects (>7 days without activity)
  • Conversion rates between stages (if enough data)
  • Signals detected this week vs last week
  • Outreach stats: sent, replied, bounced, meetings booked

Telegram format:

Pipeline Report [date]:
[N] total | [N] active | [N] stale

Top prospects:
1. Company (score) — stage — next action
2. ...

This week: [N] signals, [N] outreach, [N] replies
Conversion: signal→meeting [X]%

FIT+INTENT+TIMING Scoring

FIT (0-40) — ICP Match

FactorPointsCriteria
Company size0-10Score against the size bands in your ICP from MEMORY.md (e.g. sweet spot=10, adjacent=5, edge=2)
Use-case complexity0-10Score against the complexity profile that matches your wedge (e.g. high=10, medium=7, low=3)
Tool maturity0-10Active in the category=10, evaluating=7, greenfield=3
Budget authority0-10VP+ contact=10, team lead=7, IC=3

INTENT (0-40) — Active Buying Signals

SignalPoints
Pain language (public)8-10
Competitor evaluation8-10
Active hiring (AI/DevTools)6-8
Tech stack changes6-8
Published content on pain5-6
Event attendance3-5

TIMING (0-20) — Window Open

FactorPoints
Funding (<90 days)8-10
Hiring surge5-8
Leadership change5-7
Quarter start3-5
Signal freshness (<7d)5

Score Routing

ScoreAction
75-100CEO direct outreach, personalized demo
50-74Agent drafts, CEO approves
25-49Nurture sequence
0-24Monitor, re-score monthly

Self-Optimization (Weekly Review)

During weekly pipeline-review:

  1. Calculate conversion rates by stage
  2. Identify which signal types → highest reply rates
  3. Identify which outreach templates → highest reply rates
  4. Flag patterns: "Funding signal + CTO contact → 3x better than hiring signal + IC"
  5. Recommend weight adjustments for next week
  6. Log insights via structured-log, promote to MEMORY.md if confirmed 2+ times

Related Skills

  • signal-scanner — detects signals that feed into pipeline
  • outreach-sender — executes outreach from pipeline
  • meeting-prep — prepares for meetings with pipeline prospects
  • prospect-demo — generates [YOUR_PRODUCT] demos for prospects
  • founder-sales — sales strategy and discovery framework
  • sales-qualification — qualification framework
  • structured-log — logging pipeline events
  • hypothesis-tracker — tracking outreach template hypotheses

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.