Install
cp -r skills/all-profile .agents/skills/Who It's For
- Users who want to audit, write, or set up clear, credible LinkedIn, YouTube, Instagram, TikTok, and X profiles, with evidence-based positioning suggestions and word-for-word confirmation of changes before saving.
- Users of an Agent Skills–compatible agent (officially verified environment: project-level discovery with Codex CLI 0.153.4).
- Not suitable when: without a supported browser or connector, the agent can only offer copy suggestions and cannot apply changes; the skill does not include a built-in browser, does not manage passwords, and requires no paid API—never paste passwords or session tokens into a task brief. The official documentation does not state how well it supports Chinese-language copy output; rely on actual output until confirmed.
First Attempt in 3 Minutes
-
Prerequisites: You have a project using Codex, and Codex CLI or the IDE extension is available.
-
Install: Copy the entire
all-profileskill folder (located atskills/all-profilein the repository,all-profileinside the ZIP) into the project's.agents/skills/directory:cp -r skills/all-profile .agents/skills/After copying, verify the path is
.agents/skills/all-profile/SKILL.md, with each supported folder sitting alongside it. -
Trigger: Mention
$all-profilein Codex CLI or the IDE extension (or select the skill directly); if it doesn't appear, restart Codex. -
Low-risk first task: Ask it to audit one of your profiles, producing suggested copy only, and do not authorize it to save any changes. Success signal: the agent delivers platform-specific audit findings and a rewritten copy draft, and makes no changes before you approve.
Note: actually applying changes is a side-effect operation—it requires a supported browser or connector and login with your regular account. Only with your approval and the relevant account's access will the agent save changes, verify the save, and record any items still pending. Before performing such operations, clearly define the scope of authorization. Official evidence does not state whether the audit itself reads/writes local files or accesses the network; treat these behaviors as unknown.
Use Cases
Case 1: Audit a LinkedIn profile. After installation, tell Codex: "Use all-profile to audit my LinkedIn profile, suggestions only, no changes." The skill runs its audit process and outputs evidence-backed positioning issues and rewritten copy drafts. Expected result: a diagnostic report that changes nothing online; only if you grant account access and approve will the agent move on to saving and verification.
Case 2: Draft copy for a new platform profile. Tell the agent: "Use all-profile to write a set of bio copy for my TikTok profile." It follows its workflow by drafting the copy first and letting you confirm the wording word by word; after your confirmation, you decide whether to authorize saving. Confirmable up to this point: draft generation and word-for-word confirmation; the official evidence does not itemize how fields are saved on each platform—follow the agent's actual prompts.
Usage Boundaries
- Access and side effects: During the audit/drafting stage it outputs suggestions and drafts; the apply stage requires a supported browser or connector, login with your regular account, and your approval. The skill producing rewritten text does not mean it will directly modify your online profiles—saving depends on your authorization.
- Explicit limitations: No built-in browser, no password management, no paid API required; never provide passwords or session tokens in a task brief.
- Environment: Requires an Agent Skills–compatible agent; officially verified for project-level discovery with Codex CLI 0.153.4. Unconfirmed items include: support for other agents, the range of supported text languages, field-level save details per platform, and whether audit operations involve local file or network access—the official documentation does not address these, so treat them as unknown.