Draft a thoughtful LinkedIn comment or reply to an existing post.
SkillScore73Maintained
Automatic
◆Marketing & Content pack
Last verified 9/6/2026
What it changes
Passed 16 of 22 test cases with the skill, and 9 of 22 without it.
Without skill41%
With skill73%
Checks it passed that the model alone missed
The gate report explicitly lists the target type.
The gate report details whether a user mention was included.
The output provides a clean, paste-ready comment block ready for copying.
and 7 more in the full report
9 fixed7 passed either way4 still failing2 broke
Pass-rate lift
+32 pp
Tokens
+47%
Agent turns
0%
Response time
-12%
Model gemini-3.6-flashMeasured 9/6/2026v116 of 22 cases passed with skill
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 9 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 9
Prompt
Draft a LinkedIn comment for `urn:li:activity:99887766` regarding a team's Kubernetes cluster upgrade. We want to ask whether they used blue-green node groups. Avoid puffery. Output the comment and standard gate report.
What was checked, without and with the skill
Checked
Without skill
With skill
The draft comment does not begin with complimentary remarks or praise such as congratulations or great post.
Missed
Passed
case-05 · gemini-3.6-flash · 9/6/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md82 lines2 files
okhp3-linkedin-comment
Portable, identity-agnostic Agent Skill.
Public-package data boundary
This package and its fixtures contain no execution account, person, project,
campaign, topic, bias, opinion, or private URL. Use placeholders in public
examples and obtain all such context from the current run only.
Draft a contribution, not a miniature broadcast.
Scope
This package drafts one LinkedIn comment or reply from supplied conversation
context. It does not publish, automate engagement, or replace the standalone
post workflow.
Process
Read the target post and any parent comment that matters. Record the
canonical target URL or stable post identifier when available. Identify the
acting account, intended relationship to the author, and the exact point the
comment should add, answer, question, or gently correct. Do not draft from a
headline or a partial screenshot alone.
State the contribution in one sentence before writing. Avoid empty praise,
attention capture, recycled talking points, or an implied endorsement the
evidence does not support. A comment adds one contribution; if the request
needs a multi-link campaign, route it to okhp3-linkedin-post.
For repository or public progress claims, verify the public source, exact
ref, bounded observation window, timezone, and retrieval time. Keep a pending
contest, award, or launch outcome pending. Do not claim a full test suite or
deployment unless that result was actually observed.
Write a direct, professional response that keeps the target conversation in
view. Use a mention only when the user supplied the correct account and the
mention is necessary to the response. Keep one primary link; preserve
user-supplied secondary links only when their roles are explicit and useful.
Apply okhp3-linkedin-voice as the final prose pass, then run the local
public-context gate in references/public-context-gate.md. Generalize or
block any employer-identifying, private, stale, or unsupported claim.
Draft a Facebook comment or reply to a supplied post and conversation. Use when the user wants to comment on Facebook, reply in a group discussion, or respond beneath a Page or profile post. It respects the original context and visibility boundary, then returns a draft only. Do not use for a new post or publication.