Connections Optimizer

1---2name: connections-optimizer3description: Reorganize the user's X and LinkedIn network with review-first pruning, add/follow recommendations, and channel-specific warm outreach drafted in the user's real voice. Use when the user wants to clean up following lists, grow toward current priorities, or rebalance a social graph around higher-signal relationships.4origin: ECC5---6 7# Connections Optimizer8 9Reorganize the user's network instead of treating outbound as a one-way prospecting list.10 11This skill handles:12 13- X following cleanup and expansion14- LinkedIn follow and connection analysis15- review-first prune queues16- add and follow recommendations17- warm-path identification18- Apple Mail, X DM, and LinkedIn draft generation in the user's real voice19 20## When to Activate21 22- the user wants to prune their X following23- the user wants to rebalance who they follow or stay connected to24- the user says "clean up my network", "who should I unfollow", "who should I follow", "who should I reconnect with"25- outreach quality depends on network structure, not just cold list generation26 27## Required Inputs28 29Collect or infer:30 31- current priorities and active work32- target roles, industries, geos, or ecosystems33- platform selection: X, LinkedIn, or both34- do-not-touch list35- mode: `light-pass`, `default`, or `aggressive`36 37If the user does not specify a mode, use `default`.38 39## Tool Requirements40 41### Preferred42 43- `x-api` for X graph inspection and recent activity44- `lead-intelligence` for target discovery and warm-path ranking45- `social-graph-ranker` when the user wants bridge value scored independently of the broader lead workflow46- Exa / deep research for person and company enrichment47- `brand-voice` before drafting outbound48 49### Fallbacks50 51- browser control for LinkedIn analysis and drafting52- browser control for X if API coverage is constrained53- Apple Mail or Mail.app drafting via desktop automation when email is the right channel54 55## Safety Defaults56 57- default is review-first, never blind auto-pruning58- X: prune only accounts the user follows, never followers59- LinkedIn: treat 1st-degree connection removal as manual-review-first60- do not auto-send DMs, invites, or emails61- emit a ranked action plan and drafts before any apply step62 63## Platform Rules64 65### X66 67- mutuals are stickier than one-way follows68- non-follow-backs can be pruned more aggressively69- heavily inactive or disappeared accounts should surface quickly70- engagement, signal quality, and bridge value matter more than raw follower count71 72### LinkedIn73 74- API-first if the user actually has LinkedIn API access75- browser workflow must work when API access is missing76- distinguish outbound follows from accepted 1st-degree connections77- outbound follows can be pruned more freely78- accepted 1st-degree connections should default to review, not auto-remove79 80## Modes81 82### `light-pass`83 84- prune only high-confidence low-value one-way follows85- surface the rest for review86- generate a small add/follow list87 88### `default`89 90- balanced prune queue91- balanced keep list92- ranked add/follow queue93- draft warm intros or direct outreach where useful94 95### `aggressive`96 97- larger prune queue98- lower tolerance for stale non-follow-backs99- still review-gated before apply100 101## Scoring Model102 103Use these positive signals:104 105- reciprocity106- recent activity107- alignment to current priorities108- network bridge value109- role relevance110- real engagement history111- recent presence and responsiveness112 113Use these negative signals:114 115- disappeared or abandoned account116- stale one-way follow117- off-priority topic cluster118- low-value noise119- repeated non-response120- no follow-back when many better replacements exist121 122Mutuals and real warm-path bridges should be penalized less aggressively than one-way follows.123 124## Workflow125 1261. Capture priorities, do-not-touch constraints, and selected platforms.1272. Pull the current following / connection inventory.1283. Score prune candidates with explicit reasons.1294. Score keep candidates with explicit reasons.1305. Use `lead-intelligence` plus research surfaces to rank expansion candidates.1316. Match the right channel:132   - X DM for warm, fast social touch points133   - LinkedIn message for professional graph adjacency134   - Apple Mail draft for higher-context intros or outreach1357. Run `brand-voice` before drafting messages.1368. Return a review pack before any apply step.137 138## Review Pack Format139 140```text141CONNECTIONS OPTIMIZER REPORT142============================143 144Mode:145Platforms:146Priority Set:147 148Prune Queue149- handle / profile150  reason:151  confidence:152  action:153 154Review Queue155- handle / profile156  reason:157  risk:158 159Keep / Protect160- handle / profile161  bridge value:162 163Add / Follow Targets164- person165  why now:166  warm path:167  preferred channel:168 169Drafts170- X DM:171- LinkedIn:172- Apple Mail:173```174 175## Outbound Rules176 177- Default email path is Apple Mail / Mail.app draft creation.178- Do not send automatically.179- Choose the channel based on warmth, relevance, and context depth.180- Do not force a DM when an email or no outreach is the right move.181- Drafts should sound like the user, not like automated sales copy.182 183## Related Skills184 185- `brand-voice` for the reusable voice profile186- `social-graph-ranker` for the standalone bridge-scoring and warm-path math187- `lead-intelligence` for weighted target and warm-path discovery188- `x-api` for X graph access, drafting, and optional apply flows189- `content-engine` when the user also wants public launch content around network moves