crewAIInc / crewAIInc/crewAI

[FEATURE] Tool for stamping agent outputs to a transparency log

Open
#7,561 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
58.8k
Forks
8.5k
Avg merge
1d 15h
Merged PRs (30d)
109

Description

Feature Area

Integration with external tools

Is your feature request related to a an existing bug? Please link it here.

NA

Describe the solution you'd like

A tool that takes text an agent produced and records a hash of it in a public append-only log, returning a Merkle root and a verify URL anyone can check without an account.

The use case is an agent whose output is later disputed — a decision, an approval, a generated document — where you need to show what the output was at the time it was produced, rather than re-asking the model. Verification is a plain GET, so it works for someone who was never party to the run.

Describe alternatives you've considered

Writing outputs to application logs, which the operator can edit; commercial notary APIs, which require an account to verify; hashing locally, which proves nothing about when.

Additional context

#6388 implements this as MarkovianStampTool and is written and tested (5 passing tests, ruff clean). It was closed by the first-time-contributor bot for having no linked issue. I'll reopen it against this issue with a corrected README — the current one conflates the per-item timestamp with the protocol's chain-level Bitcoin anchor, which are separate guarantees. The tool states plainly that it proves data existed, not that it is correct.

Willingness to Contribute

Yes, I'd be happy to submit a pull request

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start with the implementation and tests referenced in issue #6388, then review its corrected README. Check how the tool records output hashes, returns the Merkle root and verify URL, and distinguishes per-item timestamps from the chain-level Bitcoin anchor. Done means the documented guarantees are accurate and the existing five tests remain passing with ruff clean.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
ai, tooling
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Active
Clarity
Mostly clear
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.