cop
Optimize agent-to-agent communication and internal scratchpads using a high-density compressed protocol.
Free
Works with the AI Agents you already use
cop
Example session with this skill installed
Initialize /cop and generate a T2 status report for a subagent regarding the broken auth-service fix.
- Read your context and instructions
- Compiled the cop
I have initialized the Compressed Communication Protocol. The following report uses T2 density for agent-to-agent handoff, prioritizing logical constraints and evidence.
COP T2 Status Report: auth-service
WIP|auth-service:v2|fix-in-progress|obs:log:402:invalid-token-header|VRF:middleware:L42
Context & Logic
- GOAL: fix 401/403 leakage on /api/v1/login.
- WHY: token-parser fails IF header-format NOT 'Bearer <token>'.
- OBS:
services/auth/middleware.go:88lacks regex validation. - INF: malformed headers crash the parser thread.
- DONE: unit-test scaffolding in
tests/auth_test.go. - NI: prod-log SHA for recent exploit attempt.
- NEXT: apply regex fix -> VRF via
go test ./....
Constraints
- ONLY modify
middleware.go. - NOT authorized for deployment.
- BLOCK(db-latency) IF test-suite exceeds 30s.
Next steps
- Review
services/auth/middleware.go:88-102for logic gaps. - Execute
make test-authto confirm the regression. - Provide prod-log SHA if unauthorized access persists.
Connects securely to your tools. The creator never sees your data.
About this skill
Stop wasting tokens on filler words and grammar when communicating between agents or writing internal scratchpads. This skill implements the Compressed Communication Protocol (COP), a system for high-density, machine-optimized text that strips away linguistic overhead without losing logical precision. It allows your agent to think, plan, and collaborate using tiered density levels, ensuring that subagent handoffs are concise while human-facing output remains professional and clear.
What it does
- Tiered Density Management applies specific compression levels (T0 to T3) based on whether the reader is a human, the agent's own memory, or a peer agent.
- Logical Preservation ensures that while articles and copulas are dropped, critical reasoning markers like NOT, IF, and BECAUSE are retained.
- Status Coding uses standardized shorthands like ACK, WIP, and BLOCK to communicate task state instantly.
- Contextual Pointer Usage replaces redundant text copying with precise file paths, line numbers, and SHA references.
- Validated Action Guardrails forces a transition back to plain prose for high-stakes actions like deletions or deployments.
How it works
- Use the
/copcommand to initialize the protocol rules and load the full reference directory. - Tag internal thoughts or scratchpad entries with T1 for maximum density during the planning phase.
- Use the standardized line format
STATUS|target|result|evidence|nextfor all agent-to-agent messages and handoff files. - Convert COP fragments back into T0 prose automatically before presenting results to the human user.
Frameworks & tools
This skill is platform-agnostic and works with any agentic framework including Claude Code, Cursor, AutoGPT, or custom LangGraph/CrewAI implementations. It is specifically designed for LLMs to reduce context window usage in multi-agent workflows.
Why this beats prompting it yourself
Writing "be concise" is vague and often leads to the loss of critical technical details. This skill provides a rigid, structured syntax that maintains logical integrity and provides a specific vocabulary for status reporting that agents can parse reliably.
Use cases
- Drafting internal agent scratchpads to save context window tokens.
- Standardizing handoff logs between primary agents and subagents.
- Writing dense technical runbooks and specs intended only for machine reading.
- Reducing latency in agent-to-agent communication via high-density fragments.
Known limitations
Requires the agent to have access to the reference/COP.md file for full vocabulary lookups. Not suitable for user-facing documentation or public-facing code comments.
Requires access to reference/COP.md for full vocabulary. Not for user-facing text. Dense shorthand can be misinterpreted if T-levels are misapplied.
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
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 2 days ago
- Free with an Agensi account