docs(readme): answer the "i can do this with one prompt" objection#480
Open
plind-junior wants to merge 2 commits into
Open
docs(readme): answer the "i can do this with one prompt" objection#480plind-junior wants to merge 2 commits into
plind-junior wants to merge 2 commits into
Conversation
the most common reaction to vouch is that a paragraph in claude.md or a host's native auto-memory already solves it. that's largely true for recall, and pretending otherwise loses the reader — so concede it up front and draw the real line: memory is a recall problem, vouch is a write-path trust problem. adds a section after the llm-wiki framing with a side-by-side table (who can write, why believe a line, what happens when it's wrong, who changed it, n>=2 writers, what you end up reading) and an honest close: solo, one prompt is fine; once several agents or a team write to shared knowledge, the pile needs an editor, and an editor isn't promptable.
|
Important Review skippedAuto reviews are disabled on base/target branches other than the default branch. Please check the settings in the CodeRabbit UI or the ⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Pro Plus Run ID: You can disable this status message by setting the Use the checkbox below for a quick retry:
✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
github-actions
Bot
added
docs
draws the section's core claim as a picture: trust in what you read stays flat behind a review gate as writers are added, while a shared memory file decays from the second writer on — identical at one writer, which the section concedes in prose. light and dark svgs swap via a prefers-color-scheme picture element; the subtitle labels the chart illustrative, not a benchmark, and the six-row table above remains the accessible table view of the same comparison.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What changed
adds a
## "I can do this with one prompt"section to the README, right after the llm-wiki framing and before the demo video. concedes the objection honestly (recall really is easy; a memory file or a host's native auto-memory is the right first reach), then draws the line the objection misses: memory is a recall problem, vouch is a write-path trust problem — single-writer memory needs no trust model, n>=2 writers do.carries a six-row side-by-side table (who can write / why believe a line / what happens when it is wrong / who changed it / behaviour at n>=2 writers / what you end up reading) and closes with the positioning: vouch is not a better place to put memory, it is a review gate in front of one and a wiki on the other side. links to docs/review-gate.md for the full case.
Why
"you can do that with one prompt" is the most common reaction to vouch, and the README never addresses it — a reader who thinks it and finds no answer just leaves. it also mis-frames the product: leading with "durable memory" invites the comparison on recall, the one axis where a paragraph in claude.md genuinely competes. this moves the argument to the axis vouch actually owns.
What might break
docs only — no code, no method surface, no behavioural change. every claim in the table was verified against the tree before writing: citation-required is a model validator (models.py:234), sources are content-hashed (models.py:166), supersede/contradict/archive exist in lifecycle.py, the approver-must-differ rule is proposals.py:380, and the audit log is append-only. no forward-looking capability is claimed — the "runs alongside whatever your host already remembers" line describes passive capture as it ships today, not the planned import-native extractor.
VEP
not a surface change; none needed.
Tests
docs/review-gate.mdverified to exist