Consulting LinkedIn Engage

Workspace: use the selected project and its AGENTS.md; keep existing entity folders and Reality headings. Business paths below are relative to that project. Bundled resources are relative to this installed skill; sibling capabilities resolve by their installed names. Never search another private checkout for missing inputs. Use workspace identity, audience, pricing, and DESIGN.md fonts. Local _work adapters and evals are optional workspace tools, not bundled dependencies. Check presence and current help first; otherwise use an available connector for the same scoped operation. If neither exists, report that step incomplete. For missing scorers, perform the stated checks and label the result manual/unscored; never invent a numeric score or successful provider action.

Social-selling via thoughtful commenting — the safe, human-in-the-loop way. The agent decides where to engage (ranked targets + angles); the owner writes and posts every comment in his own voice. This is the deliberate opposite of auto-comment tools (slop) and engagement pods (ban risk).

Why this shape

None of the LinkedIn-native commenting tools expose an API, and unattended AI comments erode a credibility-led brand. So the agent's job is targeting + drafting angles, not posting. See integrations/linkedin/linkedin-funnel-strategy.md (commenting section).

The daily split (5 + 5)

Aim for ~5 comments on ICP posts + ~5 on peer posts per day (pattern from a top operator's system — swipe/posts/magali-dereu/ANALYSIS.md, image 46-1):

• ICP posts = your buyers' posts → warm leads in public, no cold DMs needed.
• Peer posts = creators within ±50% of your follower count → their audience is your audience, so commenting borrows reach. Find them in the Activity tab of your 20–50 ICPs and where the same names recur. Peers grow reach; ICPs grow pipeline — do both.

Steps

1. Assemble the target pool from four sources:
   • Warm engagers — recent integrations/linkedin/engagement/*-enriched.json (people who already engaged the owner's posts; they carry an icp tier from score_lead).
   • Attio Warm Leads (live) — query the Warm Leads list; these are known prospects worth nurturing.
   • Fresh ICP posts (outbound) — run python integrations/linkedin/_work/find_posts.py --query "<ICP topic>" (e.g. "music AI", "catalog valuation", "label operations") to surface recent posts by ICP voices to comment on. (Apify post-search, ~$1.50/1k; the script confirms cost first.)
   • Peer creators — ~5/day within ±50% of the owner's follower count posting on adjacent topics (the reach play above). find_posts.py by topic surfaces these too.
2. Rank. Score each target by ICP fit (reuse score_lead.score_fit; tiers A→D from positioning/icp.md) × opportunity (do they have a recent post to comment on?). Comment depth beats likes; A/B tier with a fresh post ranks highest. Drop C/D unless there's a strong reason.
3. Draft angles, not comments. For each top target produce: who (name, title, company, tier), why they fit (the ICP signal), the specific post URL to comment on, and 1–2 comment angles — a genuine point of view grounded in their post + the owner's positioning (turn scattered AI experimentation into coordinated capability). Angles are seeds; the owner writes the real comment. Bias angles toward one of 4 high-reach comment styles (from the source operator's system):
   • Show personality — a real, specific reaction, not "Great post!".
   • Don't (fully) agree — add a respectful counterpoint or nuance; tension earns impressions.
   • Add value — contribute the one thing the post missed (an example, a number, a caveat).
   • Be specific — name the exact line/claim you're responding to; "you (pain point) → here's how I fixed it".
4. Output the queue. Write integrations/linkedin/engagement/<date>-engage-queue.md — a short ranked list (cap ~5–10 targets/day) the owner can work through in one sitting. Keep it scannable.
5. Guardrails (non-negotiable).
   • Never auto-comment, never connect/DM from this skill — it produces a queue only.
   • No engagement pods, ever (Lempod-style coordinated engagement = shadow-ban risk).
   • Comments stay in the owner's voice; the angles must be specific to the post, never generic praise.
   • Cap the daily volume (~5–10 thoughtful comments) and spread it out — protect the account.
6. Close the loop. Note which targets the owner engaged so a later recoup-internal-consulting-linkedin-audience pull can see who replied/engaged back → enrich → Attio. Comments that spark a reply are the warmest possible top-of-funnel signal.

Chains with

• find_posts.py (outbound target discovery) · pull_engagement.py + enrich.py (warm engagers + ICP scores) · recoup-internal-consulting-linkedin-audience (convert engagers who reply into Attio leads) · recoup-internal-consulting-followup-sequencer for the eventual follow-up (only after a real interaction).