agent-memory spec: remove tool-exposure & contradicted rows (round-2 review)#10
Merged
Conversation
Round-2 review: the tool-exposure row (old D11) presented semantic tool retrieval as an optional decision with 'pass all tools' as an equal option, when the course teaches it as best practice; and the three contradicted rows (distance strategy, index type, toolbox k) were low-stakes levers surfaced only because the course contradicted itself. Removes all four Ledger rows and renumbers to a contiguous D1-D15. No behavior lost, only relocated: - semantic retrieval stays baked into R4 (+AC4); toolbox store still D7 - distance strategy (cosine) & index (exact at fixture scale) -> §2, with the consistency invariant kept in R15/AC17 and exact-top-k in AC3 - toolbox k=5 -> R4/§4; single-config-point kept in R15/AC17 - the contradictions themselves stay as CTX-C4/C5/C6 provenance This spec now has zero contradicted rows. DEVIATION: this hand-edit runs ahead of the generation guide, which still promotes contradicted decisions to rows. Regenerating before the guide issue lands will re-introduce these rows -- do not regenerate until then. Tracked in a follow-up issue. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Why
Round-2 reviewer feedback on the agent-memory spec's Decision Ledger:
contradictedrows (distance strategy, index type, toolbox k) were low-stakes levers surfaced only because the course contradicted itself across lessons — near-equivalent or trivial choices, not architectural decisions the learner must make at build start.What changed
Removes all four Ledger rows and renumbers the remaining rows to a contiguous D1–D15. No behavior is lost — only relocated:
No business rule or AC was deleted (16 rules / 20 ACs unchanged); the rule↔AC bijection holds. The spec now has zero contradicted rows.
This is a hand-edit that runs ahead of the guide: the guide still promotes contradicted decisions to Ledger rows, so regenerating this spec before the guide is updated will re-introduce all four rows. Do not regenerate until the follow-up guide issue lands. That issue (a stakes filter for the
contradictedcategory + howdesign-arguedrows should encode taught best-practice) will be filed and linked.🤖 Generated with Claude Code