BeHuman
v1.0.0Make AI responses more human by adding inner dialogue — a Self-Mirror consciousness loop. Use when user says 'behuman', 'like a human', 'more human', 'less A...
⭐ 0· 41·0 current·0 all-time
bySway Liu@swaylq
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
The name/description (produce human-like responses via a Self↔Mirror↔Response loop) matches the provided artifacts (SKILL.md, README, claude-code examples, evals). No unexpected binaries, hosts, or credentials are required for the core functionality.
Instruction Scope
SKILL.md and related files define a prompt-level procedure (Self, Mirror, Conscious Response) and when to trigger it. Instructions do not ask the agent to read files, system config, or unrelated environment variables; they also explicitly skip the mirror for purely technical/factual queries. The skill encourages blunt/shorter phrasing for emotional scenarios — this is a design choice (not a security incoherence) but may have safety/UX implications for vulnerable users.
Install Mechanism
This is instruction-only (no install spec, no code executed on install). The README shows a copy-to-.claude example and a conceptual clawhub install command, but there is no archive download or external installer. Low installation risk.
Credentials
The skill itself declares no required env vars or credentials (correct for a prompt-layer technique). Reference docs include an example Python wrapper that calls OpenAI's API — that implies the integrator will need an OpenAI API key, but the skill metadata does not demand any credentials. This is expected for a prompt-only skill, but if you use the sample API wrapper you must supply your own API credentials separately.
Persistence & Privilege
always is false and the skill is user-invocable only. There are no instructions to modify other skills or system-wide config. It does not request persistent privileges or background execution.
Assessment
BeHuman is coherent and low-risk from a permissions/install perspective — it's a prompt-level pattern that doesn't require secrets or installers. Before installing or enabling it, consider: (1) behavioral impact: the Mirror intentionally pushes for blunt, emotionally direct language which can be helpful but may upset vulnerable users (avoid using it for crisis situations; surface resources and encourage professional help when needed); (2) integration choices: the README shows an optional API wrapper using OpenAI — if you adopt that, you'll need to provide and secure your own API keys (the skill metadata does not request them); (3) token cost: the skill increases token usage (2x–3x depending on mode); (4) quiet mode hides the internal process but it still runs — be aware the agent will perform extra internal reasoning even when only the final text is shown. If those trade-offs are acceptable, the skill appears to do what it claims.Like a lobster shell, security has layers — review code before you run it.
latestvk971ajkd6rgk8eb4crzhkck7yd840dhg
License
MIT-0
Free to use, modify, and redistribute. No attribution required.