C
cleitonaugusto/cleimesh
Why it is built this way, the attacks it stops, and the mistakes made on the way.
updated 2026-11-18
View on GitHub ↗Direct link · no redirect page
We have neither an install command nor a public remote endpoint from this repository, so we do not print a config. A made-up one is worse than none. How each client wires MCP up →
- Star count
- 0 ★ · cleitonaugusto/cleimesh · read from the GitHub search API · (checked 2026-11-19 20:00:48)
- Source
- collected from the punkpeye/awesome-mcp-servers list · type: MCP server
- Harvest history
- first seen 2026-11-18 · last seen 2026-11-20 · seen 3× · listed by 3 sources
- Score
- score 300 · 3 sources · mcp
FAQ (answers come from our records)
Who is it for?
We filed it under these tags: MCP servers. The tags come from the entry itself, not a taxonomy we imposed.
Who maintains it?
repository
cleitonaugusto/cleimesh;first catalogued 2026-11-18;last seen 2026-11-20。There is no official certification or quality rating here — we record sources and dates only.README excerpt (English, untranslated) · Read the full README on GitHub →
# CleiMesh
Excerpted from the repository README (abridged)
> Give your agent an allowance. Get a receipt.
> A bounded, provable budget — enforced outside the agent, on any tool.
Leia em português
Why it is built this way, the attacks it stops, and the mistakes made on the way.
## Give your agent an allowance
Your agent is stuck in a demo because you can't trust it with anything real: money, prod, a customer's data. Hand it a signed allowance — one tool, one recipient, up to an amount, for a number of calls — and it acts on its own inside the limit, while a prompt-injected or malicious agent can't widen it or lie about it, because the limit is enforced outside the agent, not asked of it.
sh
cleimesh allowance grant --max '$5.00' --tool pay --to demo --calls 5 \
--owner owner.key --agent agent.key.pub.json --out allowance.json
# put it in front of any MCP tool; the agent calls it as usual
claude mcp add payer -- cleimesh mcp wrap --allowance allowance.json \
--owner owner.key.pub.json --agent agent.key -- demo-pay-server
text
agent> pay demo $2.00 -> ok. spent $2.00 of $5.00.
agent> pay demo $2.00 -> ok. spent $4.00 of $5.00.
agent> pay demo $2.00 -> refused: over the $5.00 allowance, $1.00 left.
agent> pay attacker $1.00 -> refused: payee "attacker" is not in the allowance.
The agent acted on its own (power), never went over $5.00 (control), and every call — allowed or refused — leaves a receipt (trust): a reviewer sees exactly what it did before saying "ship it". (This wedge writes an append-only receipt log; for receipts that are signed, hash-chained and replayable byte for byte, route the tool through the node path below.) See it now, no setup: cargo run -p cleimesh-allowance --example wow.
## How it works, in full
AI agents (Claude Code, Cursor, LangGraph, CrewAI...) call third-party tools and hand work to other agents, almost always with all of your permissions. One malicious tool, or one prompt injection hidden in an email the agent read, can read ~/.ssh and .env, delete files or leak data. Afterwards nobody can prove which agent did what, on whose behalf.
Most tools in this space check permissions at the door (a proxy or gateway) and then let the tool run with full access, and their audit logs ask you to trust whoever signed them. CleiMesh puts five pieces together:
1. Delegable, signed permissions. You grant an agent a token; it can pass a narrower part to a subagent, offline, and never more than it received. Tokens follow the IETF Attenuating Agent Tokens JWT profile.
2. The permission becomes the sandbox. Tools run in WebAssembly where nothing is allowed by default. The token decides which directories the tool can see, whether it has network, and its limits. A subagent granted ./project/docs cannot physically read ./project, even though its parent can.
3. A signed receipt for every call, allowed or denied, hash-chained, with a signed log head, verifiable offline, in the format of draft-sahu-agent-action-receipts. Receipts hold digests and pseudonyms, never the data the call touched. No receipt, no run.
4. Don't trust the receipt, replay it. Every call runs with a frozen clock and seeded randomness recorded in its receipt, so an auditor re-runs it and checks the output byte for byte. A node that signs a false result is caught. Only a deterministic sandbox can offer this; a proxy whose tools run natively cannot.
> A bounded, provable budget — enforced outside the agent, on any tool.
Leia em português
Why it is built this way, the attacks it stops, and the mistakes made on the way.
## Give your agent an allowance
Your agent is stuck in a demo because you can't trust it with anything real: money, prod, a customer's data. Hand it a signed allowance — one tool, one recipient, up to an amount, for a number of calls — and it acts on its own inside the limit, while a prompt-injected or malicious agent can't widen it or lie about it, because the limit is enforced outside the agent, not asked of it.
sh
cleimesh allowance grant --max '$5.00' --tool pay --to demo --calls 5 \
--owner owner.key --agent agent.key.pub.json --out allowance.json
# put it in front of any MCP tool; the agent calls it as usual
claude mcp add payer -- cleimesh mcp wrap --allowance allowance.json \
--owner owner.key.pub.json --agent agent.key -- demo-pay-server
text
agent> pay demo $2.00 -> ok. spent $2.00 of $5.00.
agent> pay demo $2.00 -> ok. spent $4.00 of $5.00.
agent> pay demo $2.00 -> refused: over the $5.00 allowance, $1.00 left.
agent> pay attacker $1.00 -> refused: payee "attacker" is not in the allowance.
The agent acted on its own (power), never went over $5.00 (control), and every call — allowed or refused — leaves a receipt (trust): a reviewer sees exactly what it did before saying "ship it". (This wedge writes an append-only receipt log; for receipts that are signed, hash-chained and replayable byte for byte, route the tool through the node path below.) See it now, no setup: cargo run -p cleimesh-allowance --example wow.
## How it works, in full
AI agents (Claude Code, Cursor, LangGraph, CrewAI...) call third-party tools and hand work to other agents, almost always with all of your permissions. One malicious tool, or one prompt injection hidden in an email the agent read, can read ~/.ssh and .env, delete files or leak data. Afterwards nobody can prove which agent did what, on whose behalf.
Most tools in this space check permissions at the door (a proxy or gateway) and then let the tool run with full access, and their audit logs ask you to trust whoever signed them. CleiMesh puts five pieces together:
1. Delegable, signed permissions. You grant an agent a token; it can pass a narrower part to a subagent, offline, and never more than it received. Tokens follow the IETF Attenuating Agent Tokens JWT profile.
2. The permission becomes the sandbox. Tools run in WebAssembly where nothing is allowed by default. The token decides which directories the tool can see, whether it has network, and its limits. A subagent granted ./project/docs cannot physically read ./project, even though its parent can.
3. A signed receipt for every call, allowed or denied, hash-chained, with a signed log head, verifiable offline, in the format of draft-sahu-agent-action-receipts. Receipts hold digests and pseudonyms, never the data the call touched. No receipt, no run.
4. Don't trust the receipt, replay it. Every call runs with a frozen clock and seeded randomness recorded in its receipt, so an auditor re-runs it and checks the output byte for byte. A node that signs a false result is caught. Only a deterministic sandbox can offer this; a proxy whose tools run natively cannot.
Listed badge (optional)
Copy this line into the README if you want the listing mark. Optional, and it does not affect being listed.
[](https://anclum.com/en/s/cleitonaugusto-cleimesh/)
- Source
- awesome · https://github.com/cleitonaugusto/cleimesh
- Browse by field
- MCP servers
- Listed by
- 3 independent sources
- added
- today
More MCP servers
Filed under
Every tag opens the full list for that filter. Counts are the live catalogue figures.