Create NEXT-EASY-COMMANDS.md

Author: japertechnology

NEXT-EASY-COMMANDS.md

@@ -0,0 +1,57 @@
+# Implementable Commands (The "No-Undo" Set)
+
+This document lists commands that can be implemented in Hermes Chat that do not require a "Global Undo" (as undoing system-level side-effects like `apt install` is impossible). These commands focus on **reversible**, **idempotent**, or **contextual** operations.
+
+## 🛠 Workspace & Code Operations
+*Manipulate the codebase and inspect files.*
+
+*   `cat <path>`: View file content with line numbers.
+*   `md <path>`: Render a Markdown file in the terminal.
+*   `grep <pattern>`: Search for text/regex across the repository.
+*   `find <pattern>`: Locate files matching a glob pattern.
+*   `ls`: List directory contents.
+*   `git <command>`: Run Git commands (e.g., `git status`, `git diff`).
+*   `diff <path>`: Show changes for a specific file.
+*   `run <command>`: Execute a shell command (with a built-in safety timeout).
+
+## 🧠 Intelligence & Memory
+*Manage the agent's understanding and context.*
+
+*   `remember <text>`: Persistently save a fact or rule to `memory.log`.
+*   `memories [term]`: Search long-term memory for a specific term.
+*   `context`: View the current files and snippets held in the agent's active context window.
+*   `pin <path>`: Keep a file permanently in the agent's context to prevent it from "forgetting."
+*   `forget <id|term>`: Surgically remove specific information or assumptions from the current conversation history.
+*   `semantic-search <term>`: Find conceptually related items via local embeddings.
+
+## 🧵 Session & Threading
+*Organize and navigate conversations.*
+
+*   `new [name]`: Create and switch to a new conversation thread.
+*   `switch <id|alias>`: Switch to an existing thread by ID or name.
+*   `list`: List all existing conversation threads.
+*   `rename <name>`: Change a thread's alias.
+*   `history`: Show the conversation transcript for the current thread.
+*   `export md`: Save the current thread as a Markdown file.
+*   `retry`: Re-send the last prompt in the current thread.
+*   `again`: Re-send the last prompt in a brand new thread (for comparison).
+*   `best-of <n>`: Send the last prompt to `n` parallel threads and compare results.
+*   `multiline`: Enter a multi-line input mode for large code blocks.
+
+## ⚙️ Configuration & System
+*Adjust the agent's runtime settings.*
+
+*   `status`: View current runtime, model, thread, and git status.
+*   `model <name>`: Switch the LLM model.
+*   `provider <name>`: Switch the LLM provider (e.g., switching from cloud to local).
+*   `time`: Toggle the display of execution time.
+*   `verbose`: Toggle detailed JSONL event logging for debugging.
+*   `auto-retry [on|off|N]`: Configure the number of automatic retries for failed turns.
+
+## 📊 Telemetry & Observability
+*Monitor cost, performance, and efficiency.*
+
+*   `usage`: Show token consumption (Prompt/Completion) and estimated cost.
+*   `stats`: View lifetime project statistics (total threads, total turns, total words).
+*   `latency`: Monitor response speed (Time to First Token and total response time).
+*   `efficiency`: Calculate the ratio of meaningful content vs. overhead/metadata.