Conversation
Proposes three coordinated canon edits: 1. Add Bide as the sixth B to canon/methods/borrow-bend-break-beget-build.md 2. Create canon/constraints/borrow-evaluation-before-implementation.md (tier 1) 3. Add a Borrow-Evaluation bullet to canon/bootstrap/model-operating-contract.md Status: proposed. Per AGENTS.md, canon edits are operator-executed after acceptance — this PR contains only the proposal. Evidence: 5 independent failure cases (MCP server projects bypassing @modelcontextprotocol/sdk for a custom JSON-RPC handroll) + 2 success cases (oddkit write layer adopted; agent-messaging-service inspected and rejected, justifying minimal Build). Rationale draws on Goldratt (The Goal): local-maximum optimization at the agent station destroys global throughput at the operator-attention bottleneck.
Canon Quality —
|
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Proposes a planning-mode Borrow Evaluation as agent-binding governance, adds Bide as the sixth B to the existing 5B method, and surfaces reversibility as a load-bearing planning-time criterion.
What this PR contains
Just the promotion artifact:
docs/promotions/P0002-borrow-evaluation-before-implementation.md. No canon files are modified — perAGENTS.md, those edits are operator-executed after acceptance.What this PR proposes (for review)
Three coordinated canon edits, all delivered as working drafts alongside this proposal so the actual text is reviewable now, not described later:
canon/methods/borrow-bend-break-beget-build.md. New subsection between Beget and Build with three resolution paths:inspected-and-adopted,inspected-and-rejected,waiting(with required tripwire). Both fork outcomes framed as success.canon/constraints/borrow-evaluation-before-implementation.md. Falsifiable 6B Evaluation table the agent must produce in planning before execution. Includes inspection criteria (vision conflict, foundational gap, gross overcomplication, opinionated stack imposition, improper authority, persistent gap), Reversibility Note format, Goldratt frame, and worked examples.canon/bootstrap/model-operating-contract.md. One-line addition under Working Principles so the rule loads on first turn.Evidence
Failure cases (5): translation-helps-MCP, oddkit (MCP transport), aquifer-mcp, ptxprint-mcp, app-builder-mcp — each bypassed
@modelcontextprotocol/sdkfor a custom JSON-RPC handroll, no Borrow row, no inspection.Success cases (2): oddkit write layer (
Bide → inspected-and-adopted, Jan–Feb 2026, bet on connector pattern, humans as manual wire/bus during wait, GitHub MCP server became native connector); agent-messaging-service (Borrow → Bend → Break → Bide → inspected-and-rejected → Build = minimal, field's surfaced primitives failed inspection on multiple criteria).Rationale
Goldratt's The Goal applied to AI-augmented development: optimizing the local agent station (write code immediately, looks like progress) destroys throughput at the actual constraint (operator attention spent debugging reinvented bugs). The Borrow Evaluation is a pacing rule that keeps the bottleneck fed.
What happens after merge
If accepted, the operator executes the three canon edits in a separate PR using the working drafts shared in the conversation thread. Status moves
proposed → executed.Risk
Medium — adds a step to a stable canonical method, adds a tier-1 constraint binding all agent implementation work, introduces reversibility as a new planning-time consideration. Mitigations detailed in the artifact.
Naming open question
"Bide" is proposed as the sixth B (single B-verb meaning strategic patience). The name is the operator's call; position, semantics, three resolution paths, inspection criteria, and reversibility consideration are what matter.
Note
Low Risk
Adds a single promotion artifact (documentation only) and does not change any executable code or canon files; the main risk is process/expectation drift if the proposal is misinterpreted as an enforced rule before acceptance.
Overview
Adds a new promotion artifact
docs/promotions/P0002-borrow-evaluation-before-implementation.mdproposing a planning-mode 6B “Borrow Evaluation” gate before implementation, including introducing “Bide” as a sixth B and requiring a one-line reversibility note.The document records evidence from repeated handrolled-implementation failures, outlines three coordinated follow-on canon edits (method update, new tier-1 constraint, bootstrap bullet), and includes a risk/mitigation section—but this PR itself only adds the proposal doc and makes no canon changes.
Reviewed by Cursor Bugbot for commit 37ba9bc. Bugbot is set up for automated code reviews on this repo. Configure here.