trp2
Skill2 starsChanged 5 months agoMITView on GitHub
Run the workflow-level HIERO compaction loop — formally "Turing Recursion Protocol 2.0" (TRP 2.0). Derive new deterministic substrate components — gates, hooks, crons, skills, primitives, anchors, loops — from the real bottlenecks in OUR projects (Noesis, VibeSwap, JARVIS, Story Mode) and from the act of building them, building off existing components recursively. The hype-sphere calls this "loop engineering"; the rigorous description is workflow-level HIERO compaction + bottleneck dissolution. Invoke when Will says "run trp2", "run the learning loop", "derive components", "keep building the shell", or wants autonomous substrate self-improvement grounded in our actual work.
In this repository7 other files it gives its agents.
- classifySkill
- critical-qaSkill
- house-styleSkill
- literature-loopSkill
- marginal-contributionSkill
- md-to-pdfSkill
- ship-webSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 37
Follow existing rules; HIERO on memory writes.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
TRP 2.0 — workflow-level HIERO compaction loop (formal definition)
Compacted from its first run (2026-06-16 session): one command replaces the prompt-sequence that steered it. Per [F·compact-workflows-into-commands]. Naming: rigorous = "workflow-level HIERO compaction"; hype-reflection = "loop engineering" (cf. [F·will-original-precedes-hype-reflection]).
Args (optional)
- subject:
noesis|vibeswap|jarvis|story-mode|all(defaultall). The project whose build is the bottleneck-source this round. - n: components to derive this run (default: open-ended — keep going until told to stop or context-rotation).
Operating frame (load before deriving)
- Bottleneck lens [F·primitives-are-bottleneck-dissolutions]: every component must NAME the throughput-governing constraint it dissolves. Names no bottleneck = ceremony-noise → short-track or drop.
- Build off existing, recursively: extend rules/components we ALREADY have; do not greenfield. Each compounds the prior ([P·code-text-inspiration-loop]).
- Convergence [F·jarvis-anthropic-design-convergence]: prefer the form mirroring a harness primitive — gate↔PreToolUse/PostToolUse hook, loop↔cron/Workflow, delegation↔subagent, command↔skill, context-compaction↔HIERO.
- Honest-number [U·will-intellectual-dishonesty-frame]; pivot-until-read [F·pivot-until-read-not-global-pause] (one blocked item never freezes the loop).
Procedure (start → finish)
- SCOPE — pick subject (arg or sweep
all). State which project's build is the source. - MINE BOTTLENECKS — examine the real build (contracts, Rust, hooks, prompts, memory) for recurring friction / hard rules / silent failures. Prefer constraints already named in CLAUDE.md,
_primitives-pending.md, WAL, or session pain. - CLASSIFY (short-track, C3) — needs-structure → full component; narrow/obvious → one-line fix or log, no ceremony.
- DERIVE — build each in the convergent form:
- gate/hook →
~/.claude/hooks/*.py(PreToolUse/PostToolUse) - cron → durable
CronCreate+~/.claude/cron-prompts/*.mdcanonical (pointer+logic split) - primitive →
memory/<type>_<slug>.md(HIERO-dense) + MEMORY pointer - skill →
~/.claude/skills/<name>/SKILL.md - anchor → Schelling/reference point in the relevant doc/memory
- gate/hook →
- TEST — trust-boundary components (hooks/gates fire on every tool call) are dry-run tested against real input shapes BEFORE wiring. Never ship an untested gate; risky ones ship disabled with a note.
- WIRE — hook →
settings.json(re-validate JSON, utf-8, after); cron → durable; primitive → commit memory + pointer; skill → skills dir. A bad wire can break the session, so validate. - COMMIT — per-component. Memory → memory repo. Public-safe substrate →
jarvis-os-public/jarvis-substrate. Noesis → private remote ONLY, never public (leak-gate). - COMPACT (recursive) — if a new multi-prompt sub-workflow emerged, formalize IT into a command too ([F·compact-workflows-into-commands]).
- MENU — end with the Story Mode menu (loop continue-affordance). Keep looping until Will stops it.
Invariants
- Follow existing rules; HIERO on memory writes; keep Noesis private; one blocked item ⇒ pivot not halt.
- Each run leaves the substrate MORE operator-extension, not less ([J·jarvis-design-goal-forever]).
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

