QredenceOfficial·12 Agent Skills Included

fleet-rlm

Sandboxed recursive reasoning backend with streaming and durable sessions

Runs recursive language-model turns in isolated Daytona sandboxes with live streaming of reasoning, code, and output. Keeps sessions, attachments, artifacts, and workspace memory durable across runs without manual state management. Ships bundled skills for long-document analysis, verified data analysis, report building, and safe workspace file handling. Exposes a FastAPI plus SSE backend and terminal client so teams complete multi-step research and coding tasks faster.

51|6

↗

npx skills add Qredence/fleet-rlm --all -g -y

Available:

Gives the AI agent repository-wide execution rules, architecture invariants, and validation lanes so it modifies and runs the Fleet RLM backend and terminal client safely.

All Skills in This Repository (12)

Pure Emerald Level Indicators

📦 In Repo

Qredence

runtime-cleanup-worker

Consolidate runtime and API surfaces around fleet_rlm while preserving public contracts.

frontend-refactor-worker

Coordinate frontend file moves, import updates, and symbol renames with validation.

backend-refactor

Rewrite and clean fleet_rlm backend modules with test updates.

runtime-validation-repair-worker

Repair validator regressions across CLI, API, and browser surfaces.

optimization

Automate iterative DSPy prompt and RLM skill bundle optimization with GEPA and MLflow tracking.

dspy-programs

Design DSPy signatures and compose runtime modules for Fleet-RLM task execution.

delegation

Delegate recursive tasks to child RLM sandboxes with budget management.

browser-interaction

Fetch and inspect JavaScript-heavy web pages using a Daytona-enabled Playwright browser snapshot.

volume-bootstrap

Clarify Daytona sandbox volume layout and durable memory structure.

diagnostics

Diagnose runtime failures and observability issues in fleet-rlm using Daytona endpoints.

sandbox-execution

Execute Python code in Daytona sandboxes with durable volume persistence.

long-context

Chunk large documents and codebases for Daytona RLM workspace processing.

Official

Intermediate

Frequently Asked Questions

FAQPage Schema

How to install fleet-rlm?▼

Run `npx skills add Qredence/fleet-rlm --all -g -y` in your terminal to install all skills in this suite globally.

What does Fleet RLM do?▼

It runs DSPy recursive language-model turns inside isolated Daytona sandboxes and streams reasoning, tool calls, and code output live over SSE. Sessions, files, and artifacts stay durable across runs.

How to analyze large documents with Fleet RLM?▼

Load the long-context skill, which keeps big inputs in Python variables, retrieves only relevant excerpts, and verifies every reported fact against the original source before submitting.

Does Fleet RLM work with OpenAI-compatible providers?▼

Yes. Point the profile in config/fleet.toml at any OpenAI-compatible chat completions base URL, such as OpenAI or Databricks, and set the matching API key environment variable.

Can it generate durable reports and downloadable files?▼

Yes. The report-builder and workspace-files skills write verified reports to durable project storage and can publish them as downloadable artifacts after a successful turn commit.