agentleFS
priyanshu-aryapriyanshu-arya/LinkedIn_Automation/.claude/skills/plan-advocacy/SKILL.md

plan-advocacy

Skill0 starsChanged 2 months agoMITView on GitHub

Use when the user wants to plan or launch a team LinkedIn employee-advocacy program, define a 14-day launch plan/posting cadence/brand governance for a team, or explicitly invokes /plan-advocacy. Produces a program plan (launch schedule, per-person cadence suggestions, brand dos/don'ts, approval chain) plus a self-reported metrics template for team members to fill in -- this repo's Buffer access is scoped to the user's own channel only (REQUIREMENTS.md §12), so it cannot pull or automate analytics for any other team member's account. This is a separate module from the content-posting pipeline (research/draft/schedule skills) -- a one-off/periodic program plan and quarterly review cadence, not a recurring daily posting pipeline.

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 6 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Target content pillars: Default: offer REQUIREMENTS.md §2's 15 content pillars as a starting menu the team can narrow, rather than …

    line 38
  2. Distinctive:

    Launch date: If absent, ask, or default to "TBD — confirm before distributing."

    line 41
  3. Distinctive:

    Do not proceed past intake with a missing company name or roster — everything downstream (the launch plan, the metrics log) depends on …

    line 44
  4. Very specific:

    Every item sourced from the intake doc if one was supplied.

    line 79
  5. Distinctive:

    Never invent a number for someone who didn't submit one, and never call participation rate an "engagement" or "performance" metric in the …

    line 119
  6. Distinctive:

    Never fabricates a self-reported metric on someone's behalf: Every row in the metrics log comes from what a team member actually submitted.

    line 139
  7. Distinctive:

    Always end with: the two file paths, the confirmed program goal, and the 3-5 open decisions needing sign-off from the user or comms team.

    line 161

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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

Plan Advocacy (Employee Advocacy Agent)

This is the most standalone skill in this repo: it plans a program for a team of other people, most of whom aren't using this tool themselves. Same framing precedent as optimize-profile — this is a separate module from the content-posting pipeline (research/draft/schedule skills), a one-off/periodic program plan with a quarterly review cadence, not a recurring daily posting pipeline.

Full requirement: REQUIREMENTS.md §36. Read it if anything below is ambiguous.

Hard input contract

Same discipline as optimize-profile's "exactly two JDs": stop and ask rather than guessing if these are missing.

Required:

  • Company/team name.
  • Roster — name + role + LinkedIn profile URL per person. Used only to populate the plan document. Never fetched or scraped — per REQUIREMENTS.md §27's discipline against unofficial access to a third-party profile, which applies to a team member's profile exactly as it applies to any other person's.
  • Program goal — exactly one of: brand-awareness / hiring / thought-leadership / sales-pipeline. Don't accept a blended goal — ask which one is primary if the user names more than one.

Optional:

  • Existing brand voice/guidelines doc (path or pasted text). If absent, draft a minimal one from scratch and flag every item [DRAFTED — CONFIRM WITH LEGAL/COMMS].
  • Target content pillars. Default: offer REQUIREMENTS.md §2's 15 content pillars as a starting menu the team can narrow, rather than inventing a new pillar set for this program.
  • Launch date. If absent, ask, or default to "TBD — confirm before distributing."

Do not proceed past intake with a missing company name or roster — everything downstream (the launch plan, the metrics log) depends on having real people and a real program identity to plan around.

Deliverable format and storage

New top-level Employee-Advocacy/ folder (same precedent as Profile-Optimization/ — a periodic, re-runnable artifact accumulating history across runs, not part of the recurring content pipeline).

Two files per program run:

  1. Employee-Advocacy/YYYY-MM-DD--<company-slug>-program.md — the program plan itself, from _Templates/Employee-Advocacy-Note.md.
  2. Employee-Advocacy/YYYY-MM-DD--<company-slug>-metrics-log.md — one shared log for the whole team (never per-person files), from _Templates/Advocacy-Metrics-Log.md, that team members fill in periodically. Its program_id field cross-references the program note's id.

On a re-run for a company/team that already has a program note, treat it as an update to the existing pair (new history entry on the program note; the metrics log keeps accumulating rather than being replaced) — don't create a duplicate set of files for the same program unless the user explicitly wants a fresh relaunch.

Process

1. Intake

Per the hard input contract above. Stop and ask if company name or roster is missing. Confirm the single program goal. Ask once whether a brand-voice/guidelines doc exists; if not, say plainly that dos/don'ts will be drafted from scratch and flagged for legal/comms sign-off.

2. Draft or import brand dos-and-don'ts

Every item sourced from the intake doc if one was supplied; every item without a real source is labeled [DRAFTED — CONFIRM WITH LEGAL/COMMS]. Cover, at minimum:

  • Confidentiality — no unannounced product/financial information.
  • Tone alignment — with the company's public voice.
  • Disclosure — mention affiliation when discussing the employer.
  • No engaging with trolls/negative press — personally, on the company's behalf.
  • No political/controversial takes under the company's halo.

3. Build the 14-day launch plan

Day-by-day onboarding + first-post prompts per person. Content-angle suggestions are drawn from the pillar menu (step 1's confirmed list, or the §2 default) — never fabricate specifics about what any real named person will actually post; these are prompts/angles for them to use, not claims about content that exists.

4. Define ongoing cadence

State a recommendation (e.g. 1-2 posts/week/person) framed explicitly as a suggestion — this tool has no mechanism to enforce or schedule anything for anyone but the primary user's own channel.

5. Define the approval chain and escalation path

This tool cannot access or approve other people's drafts, so the realistic default is a spot-check pattern: advocates share drafts in a shared channel/doc during the first 14 days, tapering to self-serve after. Escalation: a named point-of-contact (comms/manager) plus a 2-step process for anything off-brand that goes live — (1) private flag, (2) takedown request. State plainly that this skill cannot itself detect or monitor that happening — there is no access anywhere in this repo to any team member's feed. Escalation is human-triggered only, always.

6. Aggregate the metrics log (re-run only)

On a run against an existing program, read whatever rows already exist in the metrics log and compute a simple rollup:

  • Most-active contributor — most self-reported entries.
  • Most-engaged post — highest self-reported numbers.
  • Participation rate — (# of roster members with at least one entry) / roster size.

Never invent a number for someone who didn't submit one, and never call participation rate an "engagement" or "performance" metric in the report — it's a completion-rate statistic about who submitted a report, nothing about how any individual post actually performed. On a first run (no existing metrics log), skip this step — the Rollup section ships saying "no entries yet."

7. Write both files and report back

File paths for both notes, plus 3-5 open decisions that genuinely need the user's or comms team's sign-off (e.g. an unconfirmed [DRAFTED — CONFIRM WITH LEGAL/COMMS] item, the launch date, the named escalation point-of-contact). Don't restate the full plan in chat — point at the files.

Hard rules

  • Never claims automated tracking of any team member's individual account performance. State the Buffer-scoping limitation (REQUIREMENTS.md §12) plainly, every time this comes up — it is a permanent architectural constraint, not a gap to be fixed later.
  • Never fabricates a self-reported metric on someone's behalf. Every row in the metrics log comes from what a team member actually submitted.
  • Never treats a missing self-report as zero — same null-handling discipline as pull-analytics. Absent means absent, not zero-performance.
  • Brand-guidance items not sourced from a real company doc are always labeled [DRAFTED — CONFIRM] (in practice, the fuller [DRAFTED — CONFIRM WITH LEGAL/COMMS]) — never presented as settled policy.
  • Never stores team members' personal contact info beyond name, role, and profile URL already given by the user.
  • Never posts, schedules, or accesses any team member's LinkedIn account. Output is a planning document only — same "produces the deliverable, human executes it" precedent as image generation (REQUIREMENTS.md §7) and Substack publishing (§25.3).
  • Never fetches or scrapes a roster member's profile. The URL is stored as given; it is never the basis for pulling that person's real content or activity.

Report back

Always end with: the two file paths, the confirmed program goal, and the 3-5 open decisions needing sign-off from the user or comms team. Make explicit that this is a planning document — the program still needs a human (the user, comms, or a manager) to actually distribute it, collect sign-off, and run the 14-day launch with the real team.

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.