Repository navigation
Conversation
`scc init` now writes RTK's usage block into CLAUDE.md/AGENTS.md whether or not the rtk binary is on PATH; the cargo build stays a question, and `--rtk` still exits 1 when the build fails (the block is written anyway). `scc update` plans and adds the block to an entry file that lacks one, keeps one already there, and takes `--no-rtk` to skip it; its JSON gains an omitempty "rtk" key in both the dry-run plan and the applied result. The update planner now treats an owned entry file that differs from its template only by the RTK and CodeGraph blocks as current, so a wired workspace reports "nothing to do" instead of listing CLAUDE.md as yours. Plan: plans/rtk-block-always.md
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
scc initwrites RTK's usage block into the entry file (CLAUDE.md/AGENTS.md) whether or notrtkis on PATH. Only--no-rtkskips it. The cargo build is still asked about;--rtkstill exits 1 when the build fails, with the block written anyway.scc updateplans a missing block alongside the template changes, adds it on apply, keeps a block already there (whatever version), and takes--no-rtkto skip.--jsongains an omitemptyrtkkey: the planned entry files under--dry-run, the spliced files on apply.internal/scaffold/update.go): an owned entry file that differs from its template only by the RTK and CodeGraph marker blocks is nowcurrent, so a wired workspace reports "nothing to do" instead of listing CLAUDE.md as "yours".Plan:
plans/rtk-block-always.md.How it was verified
scc validate --checks --pr: build, format, lint, and test (1033 tests, 86.3% coverage) all pass, 0 findings.TestInitWritesTheBlockWithoutTheBinary, the--rtkbuild failure still writes the block,TestUpdateAddsAMissingRTKBlock(dry-run text and JSON, applied JSON, second run is a no-op with a CodeGraph block present),TestUpdateKeepsAnExistingRTKBlock,TestUpdateNoRTKLeavesTheEntryFileAlone,TestPlanTreatsAnEntryFileWithOnlyUsageBlocksAsCurrent(LF and CRLF, plus a real edit staying owned).Reviews
rtkkey. Both are fixed in this PR. Tests and lint pass.AtomicWritereplaces the link with a regular file.scc rtkandscc launchalready behave this way, so it is out of scope here. Hardening it belongs inspliceEntryBlockas a separate change.