Tweet Composer

Score and optimize tweets based on X's real open-source ranking algorithm. Analyzes draft tweets against the actual ranking code — not generic tips. Use when...

This skill is instruction-only and self-contained: it reads the included rules document and returns a score/suggestions without asking for credentials or installing code. That makes it low-risk from a security perspective. Two practical notes before installing: (1) The scoring is the author's interpretation of the open-source algorithm (weights were reportedly redacted), so treat scores as heuristic recommendations rather than exact reproduction of X's ranking. (2) Ensure your agent does not have other skills or connectors that would automatically post to X on its behalf — the Tweet Composer itself doesn't post, but an agent with outbound posting permissions could use its suggestions to publish content. If you care about provenance, prefer skills with a known source/homepage; absence of a public repo or homepage is not a security problem by itself but reduces transparency.

Type: OpenClaw Skill Name: tweet-composer Version: 1.0.0 The OpenClaw skill 'tweet-composer' is benign. All files, including SKILL.md and references/algorithm-rules.md, contain instructions and content strictly aligned with the stated purpose of scoring and optimizing tweets. There are no instructions for the AI agent to perform unauthorized actions, access sensitive data, execute external commands, make suspicious network calls, or engage in any form of data exfiltration or persistence. The markdown files, treated as potential prompt injection surfaces, contain only descriptive information and operational guidance for the skill's core function.