Built for Claude Code users

The human feedback loop
for your AI workflow.

Share your Markdown files with colleagues. Get expert human feedback. Let Claude turn it into action.

Free for your first session - no credit card required

ritualdissent.com/s/a7kx92p
Files
README.md
SPEC.md
ARCHITECTURE.md
PLAN.md
SPEC.md
This service is responsible for handling all inbound webhook events from third-party payment providers. Events are queued and processed asynchronously.
Authentication
Each provider signs their payload with an HMAC-SHA256 signature. We validate this on a best-effort basis and log failures without rejecting the request. This avoids dropped events during incidents.
Retry Strategy
Failed events are retried up to 5 times with exponential backoff. After the fifth failure, the event is moved to the dead-letter queue for manual review.
Comments
R
Anonymous reviewer
This is a security risk. Signature validation must be enforced, not best-effort. If a payload fails validation, reject it - don't log and continue.
Used by teams who ship with Claude Code
No groupthink

Humans give honest feedback. Independently.

Feedback givers open the same link but never see each other's responses. No social pressure. No anchoring. Pure signal.

Anonymous by default - no account needed
Each human reviewer works in isolation
3 responses received
A
Anonymous A
"The auth section is underspecified. What happens when the token expires mid-request?"
B
Anonymous B
"Love the architecture overall. The retry logic on section 3 needs a concrete backoff number."
C
Anonymous C
"Scope is too broad for a single sprint. I'd split this into two separate specs."
Reviewers cannot see each other's responses
Precise feedback

Comment on exactly what matters.

Human reviewers highlight any section of text to leave a comment - or propose a direct edit. No ambiguity about what the feedback refers to.

Highlight any text passage
Leave a comment or suggest a replacement
Suggestions show original vs. proposed side by side
SPEC.md Highlighted passage
The service will attempt to process events within a 30-second window from receipt, after which they are considered timed out and discarded without further retry.
Suggestion
30 seconds is too short for downstream services under load. Recommend at least 5 minutes.
- within a 30-second window from receipt
+ within a 5-minute window from receipt
From feedback to action

Claude reads every comment. You decide what to do.

Run /ritualdissent feedback in Claude Code. Claude reads all responses, synthesizes them into numbered recommendations, and walks you through each one.

Feedback synthesized into concrete suggestions
Accept or reject each recommendation
Accepted suggestions applied directly to your files
Claude Code
$ /ritualdissent feedback
Fetching 3 reviewer responses...
Synthesizing feedback from SPEC.md...

Found 4 action items:
1. Enforce HMAC signature validation strictly
2 reviewers flagged the best-effort validation as a security risk
2. Extend event processing window to 5 minutes
1 reviewer suggested the 30s timeout is too aggressive
3. Specify token expiry handling in auth section
4. Add concrete backoff multiplier to retry strategy

Apply item 1? [y/n]

One Claude Skill. Real feedback.

1
/ritualdissent

Run /ritualdissent

Claude Code scans your project for Markdown files and uploads them. You get a shareable link in seconds.

2

Share the link

Send it to whoever you want feedback from. They read and comment in the browser - no account, no install.

3
/ritualdissent feedback

Run /ritualdissent feedback

Claude reads all responses and turns them into action items. You go through them one by one.

Simple pricing. No per-seat fees.

Free
0
forever
  • 1 session
  • Up to 10 Markdown files
  • Up to 10 reviewers
  • Full web reader
Get started free

Ready to find out if your spec is actually good?

Free for your first session.

Sign in with Google