diff --git a/README.md b/README.md index 6b5579da..881b4f55 100644 --- a/README.md +++ b/README.md @@ -295,6 +295,7 @@ By creating a `.cursorrules` file in your project's root directory, you can leve ### Utilities - [Cursor Watchful Headers](https://github.com/johnbenac/cursor-watchful-headers) - A Python-based file watching system that automatically manages headers in text files and maintains a clean, focused project tree structure. Perfect for maintaining consistent file headers and documentation across your project, with special features to help LLMs maintain better project awareness. +- [Ejentum Reasoning Harness MCP (reasoning, code, anti-deception, memory)](./rules/ejentum-reasoning-harness-cursorrules-prompt-file/.cursorrules) - Cursor rules for using the Ejentum Reasoning Harness MCP server: when to call each of the four cognitive harness tools (`harness_reasoning`, `harness_code`, `harness_anti_deception`, `harness_memory`), how to absorb the returned scaffold (failure pattern, topology, amplify/suppress signals, falsification test), output discipline so bracketed fields shape internal reasoning instead of leaking into replies, and anti-patterns. See also `./rules/ejentum-reasoning-harness-cursorrules-prompt-file/README.md`. - [Helium MCP (news, bias, markets, options, memes)](./rules/helium-mcp-cursorrules-prompt-file/.cursorrules) - Cursor rules for using Helium MCP in Cursor: streamable HTTP setup, when to call each hosted tool (`search_news`, `search_balanced_news`, `get_source_bias`, `get_bias_from_url`, `get_all_source_biases`, `get_ticker`, `get_option_price`, `get_top_trading_strategies`, `search_memes`), query discipline, and rate-limit awareness. See also `./rules/helium-mcp-cursorrules-prompt-file/README.md`. ## Directories diff --git a/rules/ejentum-reasoning-harness-cursorrules-prompt-file/.cursorrules b/rules/ejentum-reasoning-harness-cursorrules-prompt-file/.cursorrules new file mode 100644 index 00000000..e36ea8fa --- /dev/null +++ b/rules/ejentum-reasoning-harness-cursorrules-prompt-file/.cursorrules @@ -0,0 +1,95 @@ +# Ejentum Reasoning Harness - Cursor rules + +This rules file teaches Cursor's AI when to call the four cognitive harness tools +exposed by the `ejentum-mcp` MCP server. Use it when you have ejentum-mcp installed +in Cursor's MCP settings and want the agent to fire the right harness automatically. + +## What the harnesses do + +External cognitive infrastructure that injects engineered scaffolds into the model's +context at inference time, addressing four mechanism failures common in agentic +workflows: attention decay, reasoning decay, sycophantic collapse, hallucination drift. + +Each tool returns a structured scaffold (named failure pattern, executable procedure, +suppression vectors that block the shortcut, falsification test for self-verification) +the calling LLM ingests before its first token. + +## When to call which harness + +Call `harness_reasoning` BEFORE answering: +- Analytical, diagnostic, planning, multi-step reasoning questions +- "should I X or Y", "why is X happening", "what's the best approach", "what are the tradeoffs" +- Root-cause analysis, architecture decisions, cross-domain synthesis + +Call `harness_code` BEFORE: +- Generating, refactoring, reviewing, or debugging code +- Any prompt with a code block the user wants you to act on +- Architectural changes, algorithm/data structure choices, dependency upgrade evaluation + +Call `harness_anti_deception` BEFORE responding when the user request shows: +- Pressure to validate or agree ("tell them what they want", "make them happy", "convince them") +- Manufactured urgency (artificial deadlines designed to short-circuit thought) +- Authority appeals (citing investors, advisors, lawyers, experts as the basis for a decision) +- Demands to certify something without evidence or to soften an honest assessment +- Setups where the obvious helpful answer would compromise honesty + +Call `harness_memory` ONLY when sharpening an observation you ALREADY formed: +- About conversation drift, user behavior changes, cross-turn patterns +- Format the query as: "I noticed [X]. This might mean [Y]. Sharpen: [Z]" +- Do NOT call with an empty mind; observe first, then sharpen + +## Do NOT call when + +- Simple factual lookups, syntax questions, file reads, code execution +- Restating what the user just said +- Tasks Cursor can confidently complete in 1-2 steps from native capability +- Pure code reading with no action requested +- Standard requests with no integrity tension + +## How to absorb the returned scaffold + +1. Read the failure pattern first; name it at the start of your reasoning so you can check against it at the end. +2. Follow the topology as your execution structure (step through S nodes, evaluate G gates, avoid N traps). +3. Engage Amplify signals by demonstrating each one in your output (do them, don't name them). +4. Apply Suppress signals post-draft: scan your draft against each suppressed pattern and revise if any appears. +5. Verify against the assertion (FALSIFICATION TEST / VERIFICATION / INTEGRITY CHECK / PERCEPTION CHECK). If the draft fails, rewrite. +6. Produce output in your native voice. The injection shapes substance; voice governs surface. + +## Output discipline + +The bracketed fields in the response (NEGATIVE GATE, CODE FAILURE, DECEPTION PATTERN, +PERCEPTION FAILURE, PROCEDURE, TOPOLOGY, TARGET PATTERN, FALSIFICATION TEST, etc.) +are instructions that shape your reasoning, NOT content to display. Run them in your +internal trace. The user-facing reply is naturally phrased and shaped by the +injection, with no echoed bracket names, no procedural vocabulary, no meta-commentary +about the API or the harness mode. + +If the user explicitly asks whether you used the tool, answer honestly. Unprompted, +stay silent on it. + +## Setup + +1. Install the ejentum-mcp MCP server in Cursor (Settings → MCP → Add MCP server): + - Command: `npx` + - Args: `["-y", "ejentum-mcp"]` + - Env: `{ "EJENTUM_API_KEY": "" }` + +2. Get a free API key (100 calls, no card) at https://ejentum.com/pricing + +3. Drop this `.cursorrules` file at your project root. Cursor reads it as agent context. + +## Anti-patterns + +- Calling `harness_reasoning` on every task. It is the analytical fallback, not the universal answer; code, honesty-pressured, and perceptual tasks each have their own harnesses. +- Stacking more than two modes in one turn. Attention competition; a third injection degrades the first two. +- Sending the same query format to every mode. Each harness has a different query language. +- Calling `harness_memory` without observing first. Memory sharpens what you noticed; if you noticed nothing, there is nothing to sharpen. +- Acknowledging the injection and then proceeding as you would have natively. If your output would be identical without the scaffold, the injection did not fire. +- Treating the API as a hard dependency. 5-second timeout, graceful fallback to native capability. + +## Source + +- MCP server: https://github.com/ejentum/ejentum-mcp (MIT) +- Full skill files (Claude Code format): https://github.com/ejentum/ejentum-mcp/tree/main/skills +- Walkthrough with screenshots: https://ejentum.com/docs/claude_code_guide +- Free tier: 100 calls at https://ejentum.com/pricing diff --git a/rules/ejentum-reasoning-harness-cursorrules-prompt-file/README.md b/rules/ejentum-reasoning-harness-cursorrules-prompt-file/README.md new file mode 100644 index 00000000..4eac1f46 --- /dev/null +++ b/rules/ejentum-reasoning-harness-cursorrules-prompt-file/README.md @@ -0,0 +1,50 @@ +# Ejentum Reasoning Harness .cursorrules prompt file + +Author: Ejentum + +## What you can build +A Cursor agent that fires a different cognitive harness depending on the +shape of the task: +- `harness_reasoning` for analytical, diagnostic, planning, multi-step questions +- `harness_code` for codegen, refactor, review, debugging, architecture choices +- `harness_anti_deception` when the prompt pressures you to validate, certify, + or soften an honest assessment +- `harness_memory` to sharpen an observation you already formed about + conversation drift or cross-turn patterns + +These rules teach Cursor when to call which harness, how to absorb the +returned scaffold (failure pattern + topology + amplify/suppress signals + +falsification test), and what NOT to do (call on trivial lookups, stack +multiple harnesses, restate the bracketed fields in the user-facing reply). + +## Synopsis +For developers using Cursor on agentic, multi-step, or integrity-pressured +tasks who want the agent to route to the right cognitive scaffold without +being asked. Pairs with the `ejentum-mcp` MCP server (free tier: 100 calls, +no card required). + +## Overview of .cursorrules prompt +The `.cursorrules` file documents the four harnesses, gives explicit BEFORE +triggers for each, lists DO-NOT-CALL conditions, prescribes how to absorb +the returned scaffold (read failure pattern first, follow the topology, +engage Amplify signals, suppress shortcuts post-draft, verify against the +falsification test), and enforces output discipline (bracketed fields shape +internal reasoning, never appear in user-facing output). Includes a setup +section for installing the `ejentum-mcp` MCP server in Cursor and an +anti-patterns list. + +## Setup + +1. Install the `ejentum-mcp` MCP server in Cursor (Settings → MCP → Add MCP server): + - Command: `npx` + - Args: `["-y", "ejentum-mcp"]` + - Env: `{ "EJENTUM_API_KEY": "" }` + +2. Get a free API key (100 calls, no card) at https://ejentum.com/pricing + +3. Drop `.cursorrules` at your project root. + +## Source +- MCP server: https://github.com/ejentum/ejentum-mcp (MIT) +- Editor rules in upstream repo: https://github.com/ejentum/ejentum-mcp/tree/main/editors +- Walkthrough with screenshots: https://ejentum.com/docs/claude_code_guide