Repository navigation
docs: add a roadmap and a vision, and make the stack count CI-enforced - #28
Merged
Merged
Conversation
13 open issues, two milestones, and nothing saying which mattered or where this is going. A contributor picked by title and hoped. The vision section matters more here than in most repos because this tool writes infrastructure other people apply to real accounts, and its value is entirely in what it refuses to emit. Writing down that there will never be a flag downgrading a refusal to a warning gives a reviewer something to point at when the next PR adds --allow-public-database with a reasonable justification attached. That is the most common request a scaffolding tool receives. Now/Next/Later are tables of real linked issues mapped to the two milestones. All 13 issue references and both PR references were verified against the tracker rather than written from memory. The roadmap claims its stack count is enforced, so it is: two tests assert ROADMAP.md against the registry. Configuring a stack without updating the sentence fails with a message naming the new number. Verified by flipping s3 to configured and watching it fail: AssertionError: ROADMAP.md must say 'Four of the nine advertised stacks are configured'. The registry has 4 configured of 9; update the roadmap in the same PR that changes the code. That is the difference from websieve's roadmap, which went stale within weeks under a convention alone (websieve#34). A convention is only as good as the next person's memory. Closes #27 Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
ehtishammubarik
added a commit
that referenced
this pull request
Sep 4, 2026
The Unreleased section described four of the eight commits merged since the tag. The four it omitted are the ones a prospective contributor most needs: there is now a contributing guide, a roadmap, and a production guardrail check, and the changelog said none of that existed. - Opened a real `## [0.1.1]` heading and bumped pyproject to match, so the version the changelog releases and the version the wheel carries agree. - Added the missing entries: production intent guardrails (#5, #6), the contributing guide (#18), the roadmap with its CI-enforced stack count (#28), and the removal of the stale README test figures (#17, #20). - Removed "Twelve tests" and "64 resources to add" from Unreleased. That is #20 recurring in the one file that change did not cover. A number in a released entry is a dated snapshot and may stand; a number in Unreleased is not yet history and was simply wrong. - Replaced them with the gates that are actually enforced: the 88% coverage floor on 3.10/3.11/3.12, and the release workflow asserting on the published artifact that guardrails still refuse an 0.0.0.0/0 database and that the generated output carries no literal credential. - Stated plainly that there is no `terraform plan` in the release workflow. The eks+rds plan was run by hand, and a figure from a manual run is exactly what this entry declines to quote. Filed #31 to automate it. - ROADMAP: the same 64-resource claim, in the present tense, replaced with what CI does verify. Verified: pytest 91 passed, coverage 90.90% against the 88% floor. Closes #30 Claude-Session: https://claude.ai/code/session_012fMM2HhoJ2XzDkfMR2HwSa Co-authored-by: Claude Opus 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
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.
Closes #27. One issue, one PR.
13 open issues, two milestones, and nothing saying which of them mattered or where this is going. A contributor picked by title and hoped. That is the same guessing problem #16 solved for how to contribute, left unsolved for what.
The vision section, and why it earns its place here
This tool writes infrastructure other people apply to real cloud accounts. Its value is entirely in what it refuses to emit. So the roadmap states, permanently:
That paragraph exists so a reviewer has something to point at when a PR arrives adding
--allow-public-databasewith a reasonable-sounding justification. Without it, the refusals get eroded one exception at a time by people each individually acting in good faith.Also written down: never emit a credential in any form, never vendor
terraform-aws-modules, no drift detection or control plane, and no runtime dependency of generated repos onstackmason— you should be able to delete this tool afterwards and lose nothing.Structure
Now lists what actually works, including the uncomfortable part: three of nine stacks are configured. Next and Later are tables of real linked issues mapped to the two milestones, with the six stub stacks ordered and annotated (
s3flagged as the best first one;observabilityflagged as the largest and needing a design discussion).All 13 issue references and both PR references were verified against the tracker, not written from memory.
The part that makes this different from websieve's roadmap
websieve's roadmap went stale within weeks under a stated convention alone: two shipped features sat under Next, one labelled "the biggest correctness gap" (websieve#34). A convention is only as good as the next person's memory.
So the number here most likely to rot is asserted:
A second test asserts every unconfigured stack has a roadmap row pointing at its issue, so a stub cannot become a dead end for a contributor who has just read that it does not work.
I checked both fail before trusting them, then reverted.
Verification
terraform,kubernetes, andhelmgates skipped, not passed.Scope
One issue. The six stack implementations (#21-#26) are untouched and are what 0.2 is for.