- Home
- Skills
- Agents & Orchestration
- Endless Loop | Autonomous AI Research & Task Execution Agent
Works with the AI Agents you already use
Endless Loop | Autonomous AI Research & Task Execution Agent
A structured, stateful research and execution workflow for advancing complex, multi-step objectives.
$11.99
Endless Loop | Autonomous AI Research & Task Execution Agent
Example session with this skill installed
Research the current state of SolidJS for enterprise use. Run 3 cycles to find performance benchmarks, library ecosystem maturity, and security features. Save progress to solid-research.md.
- Read your context and instructions
- Compiled the endless loop
I have initiated the research workflow for SolidJS enterprise readiness. I completed three discovery and validation cycles, focusing on runtime performance, ecosystem compatibility with existing enterprise UI kits, and security advisories. The findings are documented in the progress record below.
Objective: SolidJS Enterprise Readiness Audit
| Metric | Status | Findings |
|---|---|---|
| Performance | Verified | 0.7ms startup time; 1.2x faster than React in DOM reconciliation benchmarks. |
| Ecosystem | Partial | Core primitives exist; UI kits like Kobalte are maturing, but legacy enterprise wrappers are missing. |
| Security | Verified | No high-severity CVEs in core; fine-grained reactivity reduces XSS surface area in templates. |
| Maturity | In Progress | Version 1.8.x is stable; migration paths for large-scale React apps remain complex. |
Cycle 1-3 Summary
Research: Investigated official SolidJS documentation, Ryan Carniato's performance deep-dives, and GitHub security advisories.
Findings: SolidJS excels in TBT (Total Blocking Time) metrics but requires specialized patterns for global state management in micro-frontends.
Verification: Benchmarks were cross-referenced between the JS Framework Benchmark and independent community audits.
- Sources:
solidjs.com,stefandoetschel.github.io/js-framework-benchmark,github.com/solidjs/solid/security.
Next steps
- Review
solid-research.mdfor the full breakdown of component-level security audits. - Approve Cycle 4 to draft a migration strategy for the specific legacy React components identified in scope.
- Provide access to internal UI requirements to compare against available Solid primitives.
Connects securely to your tools. The creator never sees your data.
About this skill
Complex research and execution tasks often stall when agents lose track of prior findings or repeat redundant work. This skill implements a rigorous, stateful workflow that advances long-horizon objectives through bounded cycles of research, execution, and verification.
What it does
- Stateful recovery reads the latest progress record to ensure every new cycle builds on verified findings rather than starting over.
- Evidence-based research gathers and cross-checks information from primary sources, recording citations and publication dates for traceability.
- Cyclical analysis identifies confirmed assumptions, remaining uncertainties, and high-value next actions to maintain momentum.
- Verified execution performs authorized actions and validates the output against success criteria before treating it as a fact.
- Durable logging maintains a structured progress record including cycle history, decisions made, and unresolved risks.
How it works
- Define objective by establishing the goal, deliverables, success criteria, and strict stop conditions.
- Execute cycles through sequential phases of discovery, validation, planning, and execution.
- Verify results by comparing tool outputs against the original success criteria to identify partial completions.
- Handoff deliverables with a final report covering verified insights, evidence sources, and remaining limitations.
Frameworks & tools
This skill is framework-agnostic and operates within any environment providing research tools, file system access for state persistence, or browsing capabilities. It relies on structured Markdown for state tracking and evidence logging.
Why this beats prompting it yourself
Standard prompts often suffer from context drift or repetitive loops during long tasks. This skill enforces a "Recover-Research-Analyze-Verify" loop that prevents the agent from claiming success without evidence or repeating failed actions.
Use cases
- Market opportunity briefs built through iterative source review and competitor analysis.
- Technical roadmaps comparing architectural approaches based on documentation and constraints.
- Codebase audits that prioritize issues and prepare verified fixes across multiple rounds.
- Project plan development that synthesizes research, feedback, and resource constraints into a timeline.
Known limitations
Does not support background execution or scheduling unless provided by the host environment. Requires an authorized toolset for external research or file modifications.
Requires host environment to support persistent file storage or session memory for state recovery. Execution is limited to user-authorized tools and cycle budgets.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 14 days ago
Permissions
No special permissions declared or detected