How do I install the Psyche Network Rules Reference skill?

Run npx skillmds@latest add lingtai-ai/psyche-network-rules-reference in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Psyche Network Rules Reference safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Psyche Network Rules Reference?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Psyche Network Rules Reference free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Psyche Network Rules Reference?

lingtai-ai (@lingtai-ai) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Psyche Network Rules Reference by lingtai-ai

Name: Psyche Network Rules Reference
Author: lingtai-ai

File contents

Psyche network-rules reference

.rules is a heartbeat signal, not a Psyche action or generic instruction API. The agent heartbeat (_check_rules_file) consumes it, unlike ordinary durable edits, which require file→context.rebuild.

Write and consume

Write the complete approved UTF-8 body atomically to the explicitly authorized agent workdir root (POSIX example; use the active shell’s equivalent elsewhere):

target='/absolute/path/to/authorized-agent'
body='/absolute/path/to/approved-rules.txt'
tmp=$(mktemp "$target/.rules.XXXXXX") &&
  cat "$body" > "$tmp" && mv "$tmp" "$target/.rules"

A failed command requires inspection of that exact temporary file; do not blindly replay or announce success. Confirm each target separately—there is no descendant broadcast. Do not substitute system/rules.md for this live signal.

The next runnable heartbeat reads and unlinks .rules before deciding whether to apply it. Read/unlink failure leaves it unconsumed and stops processing, so disappearance alone is not proof of success. A nonempty body completely replaces system/rules.md and the protected rules section; whitespace-only content is a consumed no-op. Identical content is consumed without rewrite or flush. Changed content persists and flushes, logging rules_loaded; a canonical-file write failure aborts before prompt mutation and logs rules_write_error.

Boot and every full rebuild/refresh/molt reread system/rules.md independently; a missing or empty canonical file removes the section. Identical comparison ignores leading/trailing whitespace. A pending .rules file is not yet applied, and an empty signal cannot clear existing rules. Verify the canonical file on disk and the effective protected section only after that agent's heartbeat processed the signal (or after later reconstruction).

Boundary

Writing another reachable agent's .rules file is an ordinary filesystem write, not a dedicated avatar authorization path. Reachability is not permission: only actual human scope and the target's real accessibility authorize the write. Keep this distinction visible when operating across agents.

---name: psyche-network-rules-referencedescription: Deep reference for the separate `.rules` heartbeat signal: authorized atomic writes, consumption, replacement, persistence, verification, and boundaries.---# Psyche network-rules reference`.rules` is a heartbeat signal, not a Psyche action or generic instruction API.The agent heartbeat (`_check_rules_file`) consumes it, unlike ordinary durableedits, which require file→`context.rebuild`.## Write and consumeWrite the complete approved UTF-8 body atomically to the explicitly authorizedagent workdir root (POSIX example; use the active shell’s equivalent elsewhere):```shtarget='/absolute/path/to/authorized-agent'body='/absolute/path/to/approved-rules.txt'tmp=$(mktemp "$target/.rules.XXXXXX") && cat "$body" > "$tmp" && mv "$tmp" "$target/.rules"```A failed command requires inspection of that exact temporary file; do not blindlyreplay or announce success. Confirm each target separately—there is no descendantbroadcast. Do not substitute `system/rules.md` for this live signal.The next runnable heartbeat reads and unlinks `.rules` before deciding whether toapply it. Read/unlink failure leaves it unconsumed and stops processing, so disappearancealone is not proof of success. A nonempty body completely replaces`system/rules.md` and the protected `rules` section; whitespace-only content is aconsumed no-op. Identical content is consumed without rewrite or flush. Changedcontent persists and flushes, logging `rules_loaded`; a canonical-file write failureaborts before prompt mutation and logs `rules_write_error`.Boot and every full rebuild/refresh/molt reread `system/rules.md` independently;a missing or empty canonical file removes the section. Identical comparisonignores leading/trailing whitespace. A pending `.rules` file is not yet applied,and an empty signal cannot clear existing rules.Verify the canonical file on disk and the effective protected section only afterthat agent's heartbeat processed the signal (or after later reconstruction).## BoundaryWriting another reachable agent's `.rules` file is an ordinary filesystem write,not a dedicated avatar authorization path. Reachability is not permission: onlyactual human scope and the target's real accessibility authorize the write. Keepthis distinction visible when operating across agents.

Frequently asked questions

Psyche Network Rules Reference