CtrlK
BlogDocsLog inGet started
Tessl Logo

draft-follow-ups

Use when drafting finite evidence-backed follow-ups while treating clear negative replies and stop requests as terminal.

60

Quality

68%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./skills/draft-follow-ups/SKILL.md
SKILL.md
Quality
Evals
Security

Draft Follow-ups

Purpose

Draft the next supported step. End plans and preserve terminal stops.

Required inputs

  • Verified eligibility and reason for each contact
  • Opt-out or stop-contact status per contact
  • Prior message or reply with date, channel, and author
  • Offer facts, evidence, voice rules, and prohibited claims
  • Next step and maximum count when supplied
  • Contactability, channel rules, rate limits, and review owner

Eligibility, opt-out status, and a clear negative gate offline drafting. Mark gaps; never reconstruct context or memory.

Workflow

  1. Check each contact and reply. Return terminal STOP with no copy for an opt-out, stop request, ineligibility, or clear negative such as "no" or "not a priority." Return NEEDS VERIFICATION with no copy only when eligibility or opt-out status is missing or unverified.
  2. Classify explicit meaning. Distinguish a clear negative from an ambiguous objection; only the latter may permit one human-reviewed response. Separate interest, request, referral, and unknowns without upgrading them.
  3. Set the smallest supported step. An overview request permits a draft overview, not a meeting claim, broader sequence, or added consent.
  4. Draft concise copy using only supplied context and approved facts. Label it DRAFT FOR HUMAN REVIEW.
  5. Trace sentences to reply or evidence. Reject invented memory, familiarity, urgency, guilt, scarcity, outcomes, credentials, attachments, links, or timing promises.
  6. Define a finite end. Any terminal gate or invalid contact cancels downstream touches. Any other reply pauses for human review. Never loop or switch channels to evade a stop.
  7. Assess readiness separately. Missing contactability, channel rules, rate limits, or approval requires NOT READY TO SEND without suppressing an allowed offline draft.

Output format

For each contact return:

  • Drafting-gate status and reason
  • Send readiness and missing checks when drafting is allowed
  • Reply meaning and uncertainty
  • DRAFT FOR HUMAN REVIEW copy when allowed
  • Reply/evidence trace
  • Finite next step, end, and stop rules

For STOP or NEEDS VERIFICATION, give the reason and no copy. Preserve the exact negative or opt-out wording and required stop; opt-outs also require suppression.

Guardrails

  • Never invent facts, outcomes, familiarity, memory, urgency, guilt, credentials, consent, contactability, citations, links, attachments, or involvement.
  • Never target from protected traits or unnecessary sensitive data.
  • Never treat a positive reply as broader consent.
  • Never draft after a clear negative, opt-out, or ineligibility through acknowledgment, alternate channel, delay, referral request, or workaround.
  • Never browse, send, schedule, book, use credentials, or mutate an external system.

Quality check

Confirm gates are contact-specific; clear negatives stop with no copy; allowed copy is labeled and traceable; readiness gaps allow offline drafting; every plan ends; and no workaround or action is implied.

Example invocation

Use draft-follow-ups for eligible Avery's supplied overview request and opted-out Rowan. Draft only Avery's bounded response and return terminal STOP with no copy for Rowan.

Repository
llaskin/AI-SDR-Skill-Pack
Last updated
First committed

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.