[brettheap]

Lane: openreposhape-1 (openRepoShape-1)
Claim: #521 (comment)
Claim: #427 (comment)

What

adopt-project.py and shape-doctor.py read a repository's origin one way, with one function in scripts/repo_shape.py. That reading serves #499's source check, the organisation plan fills in, what plan says about either, and the doctor's adoption line.

• A url names the <org>/<repo> it ends in, however it ends (adopt-project.py: read <org>/<repo> from every origin git accepts (a local path, a Windows backslash path, a trailing / or .git/) #521). https://github.com/acme/Atlas/, .../Atlas.git/, ssh://git@ssh.github.com:443/acme/Atlas.git/, git@github.com:acme/Atlas/ and [::1]:acme/Atlas.git/ all name acme/Atlas. So --project Atlsa --org acme is refused adopt-not-the-source, and with no --org the organisation is acme.

• A path names no repository, however it is written (adopt-project.py: read <org>/<repo> from every origin git accepts (a local path, a Windows backslash path, a trailing / or .git/) #521). This covers an absolute or relative path, / or \, a drive (C:\src\acme\Atlas.git and C:/src/acme/Atlas.git alike), a UNC share, and a file:// url. --org and --project are not compared with it, and plan says so in a NOTE that names the repository execute will act on:

  NOTE the source's `origin` is the path `C:\\src\\acme\\Atlas.git`, which names no repository on GitHub, so `--org` and `--project` were not compared with it.
       `execute` without `--local-remote-dir` pushes the split to acme/Atlsa, opens its pull request there and sets its topic on it, so that must be the source's own repository on GitHub.
  

  A path gives plan no organisation either. With no --org it is adopt-no-org, whose sentence now names this case: "...it has none, its URL is not UTF-8 or ends in no org/repo, or it is a path, which names no repository on GitHub".

• An organisation plan read and checked_value refuses is named by where it came from (adopt-project.py plan takes the organisation from the source's origin URL and runs it through checked_value("--org", ...), so an origin whose organisation part is not ASCII or holds a space (/x/café/repo.git, /x/my repos/repo.git, caf%C3%A9) ends plan in REFUSED unsafe-value: --org is 'café', a flag nobody passed, with a remediation that explains a leading - and never says to pass --org #427). The remediation says first to pass --org. The refusal and its exit code stay:

  REFUSED unsafe-value: --org (not given, so the organisation of the source's `origin`) is 'café', which is not a value this tool will put on a `git` or `gh` command line
  Remediation: pass --org <org>; without it, `plan` reads the organisation from the source's `origin`. A leading `-` is refused because git reads its own arguments (`--upload-pack=…` is a command, not a branch); the rest must match ^[A-Za-z0-9._/@+~-]{1,255}$.
  

  adopt-not-the-source names an organisation nobody passed the same way: --project Atlsa, with the organisation of the source's origin (--org not given), names acme/Atlsa, .... For --source <org>/<repo> it says "of --source".

• The doctor's adoption line reads origin the same way (shape-doctor.py, NOT A SHAPE ROOT on a git repository: print both ways in, as AGENTS.md says, and add --org to the adopt line when there is no origin #545, shape-doctor.py: a git repository is told of both ways in, the adoption and the scaffold, and the adoption line ends --org <your-org> where its origin names no organisation plan can read, so pasting it is no longer refused adopt-no-org (#545) #553): shape-doctor.py's origin_org uses the one reading, so its NEXT ends --org <your-org> for a path origin and not for a url ending /.

Why (measured on main bfed143, whose adopt-project.py main f6eaab7 still has; git 2.43.0, Python 3.12.3, local only, a stub gh)

Re-measured on main 64fc9a8, whose adopt-project.py has #567's changes, every row below gives the same result as on bfed143.

row	origin, flags	main bfed143	this PR
#521 path	<dir>/clones/Atlas.git, --org acme --project Atlas	exit 2 adopt-not-the-source ... the source is clones/Atlas; told to pass --org clones	exit 0, NOTE: not compared
#521 path, control	<dir>/srv/acme/Atlas.git, no --org, --project Atlas	exit 0, org acme read from a directory name	exit 2 adopt-no-org (pass --org)
#521 backslash	C:\src\acme\Atlas.git, --org acme --project Atlsa	exit 0, nothing said	exit 0, NOTE: not compared
#521 backslash, / twin	C:/src/acme/Atlas.git, --org acme --project Atlsa	exit 2 adopt-not-the-source	exit 0, NOTE: not compared (the same path reads the same)
#521 trailing /	https://github.com/acme/Atlas/, --org acme --project Atlsa	exit 0	exit 2 adopt-not-the-source, the source is acme/Atlas
#521 .git/	https://github.com/acme/Atlas.git/, same flags	exit 0	exit 2 adopt-not-the-source
#521 trailing /, no --org	https://github.com/acme/Atlas/, --project Atlas	exit 2 adopt-no-org	exit 0, org acme
#427 path	<dir>/x/café/repo.git, no --org	exit 2 unsafe-value: --org is 'café'	exit 2 adopt-no-org (pass --org)
#427 path	<dir>/x/my repos/repo.git, no --org	exit 2 unsafe-value: --org is 'my repos'	exit 2 adopt-no-org
#427 url	https://github.com/caf%C3%A9/repo.git, no --org	exit 2 --org is 'caf%C3%A9'	exit 2 --org (not given, so the organisation of the source's origin) is 'caf%C3%A9', pass --org
#427 url	git@h:-evil/repo.git, no --org	exit 2 --org is '-evil'	exit 2, named and --org remedy as above
#427 url	https://example.com/café/Northwind.git, no --org	exit 2 --org is 'café'	exit 2, named and --org remedy as above
#427 control	git@github.com:Org/Northwind.git, --project Northwind	exit 0	exit 0
#427 control	<dir>/y/café/repo.git, --org testorg --project Northwind	exit 2 adopt-not-the-source (#499 compared café/repo)	exit 0, NOTE
#427 no origin	no origin, no --org	exit 2 adopt-no-org (already fixed by #353, #292)	exit 2 adopt-no-org
#499 naming	https://github.com/acme/Atlas.git, --project Atlsa, no --org	--org acme and --project Atlsa name acme/Atlsa (a flag nobody passed)	--project Atlsa, with the organisation of the source's origin (--org not given), names acme/Atlsa

Only the no-origin row of #427 was already fixed on main, by #353's adopt-no-org ordering, which this PR builds on: adopt-no-org is still asked before checked_value.

How

• The decision taken: a path origin names no repository, for v1.1.0. Lane 1's coordinator accepted reviewer r521's recommendation on 2026-10-15. Brett may overrule it before the merge. The reasoning follows.

• What a local-path origin names, and why it is nothing. adopt-project.py plan never compares --project (or --org) with the repository it adopts in place, so a source named billing-service is refused with a remediation that steers to --project BillingService, which plans and checks plan ok against acme/billing-service, and execute then acts on acme/BillingService: a raw git ls-remote failure when no such repository exists, and, when a different repository of that name does, both legs, the split branch, the pull request and the topic on that repository #499 compares <org>/<project> with the source because execute without --local-remote-dir acts on https://github.com/<org>/<project>.git. Under --local-remote-dir, the split goes to the source's own path, and nothing is compared that matters. A path is a directory on this machine. It says nothing about which GitHub repository the source is.

  • Its last two directories (main): this refuses the right source whenever the path does not end in <org>/<repo> (adopt-project.py: read <org>/<repo> from every origin git accepts (a local path, a Windows backslash path, a trailing / or .git/) #521's first row). It also offers a directory's name as the organisation.
  • Its last component compared with --project only: this still refuses the right source when a working clone sits in a folder named otherwise (git clone ~/src/atlas-work). It still needs --org from somewhere.
  • No repository (chosen): this never refuses the right source. It still refuses a mistyped --project or --org wherever the origin really names a repository: a url, or --source <org>/<repo>. The cost is that a typo under a path origin plans, so the NOTE says that nothing was compared and names the repository execute will act on.

  Following the path to that repository's own origin would keep more comparisons. It would also read another repository's config, and its remote-tracking branches are not GitHub's (_places_check_asks). It is not done here.

• _origin_url is the one read of the url (git remote get-url origin, as before). _remote_repository and the NOTE's _origin_path both use it. Before, _origin_path read remote.origin.url raw through _config_value.

• repo_shape.url_repository returns None for repo_shape.origin_is_a_path (remote_is_a_path, or a file:// url). Otherwise it takes the url's path: after <scheme>://<host>/, after git's scp colon, or after ]: for a bracketed host. It drops trailing / and one .git, and takes the last two segments, with case kept.

• ONE definition, in scripts/repo_shape.py, imported by adopt-project.py and shape-doctor.py. shape-doctor.py: a git repository is told of both ways in, the adoption and the scaffold, and the adoption line ends --org <your-org> where its origin names no organisation plan can read, so pasting it is no longer refused adopt-no-org (#545) #553 (b4c1f50) had given the doctor's origin_org a copy of the regex this PR replaces, held to plan by a parity test. On the merge with main that copy and plan disagreed (r521's F1), so the copy, _ORIGIN_REPOSITORY_RE, is deleted. Neither tool is copied into a project, and no template file imports the new names, so update-shape.py apply --accept-local scripts/repo_shape.py can keep an old helper that a newly copied file imports from, and make bootstrap then fails with ImportError #536's kept-helper rule (update-accept-local-missing-name, bbcfd34) is not touched.

• Main 64fc9a8 is merged into the branch as its own commit (2cfbe0e, clean). The doctor change follows in a separate commit.

• _planned_org replaces the org block in _checked_plan_inputs. That function stays no more complex than Sonar allows; its block is now two lines. _refuse_another_repository takes org_given. _org_read_from says "the source's origin" or "--source".

• With no repository read from a path, the path branch of _rename_first and _own_name_remediation (adopt-project.py: --org and --project must name the repository the source names, compared without case, so plan refuses --project BillingService for acme/billing-service as adopt-not-the-source instead of writing a plan whose execute acts on acme/BillingService, check finds and execute refuses a plan edited to name another as plan-not-the-source, and a source whose own name the naming policy refuses is told to rename it first, never to retype --project (#499) #513's review, NIT-3) is unreachable, so it is removed. Rename-first always means on GitHub.

Tests

tests/test_adopt_origin_repository.py: 28 new tests (9 functions). All 28 fail with the three changed files at main 64fc9a8 (that run fails 37: the 28 new tests, the 6 changed existing ones, and the doctor's [trailing-slash], [posix-path] and [file-url]).

Existing tests changed because this PR reverses what they asserted:

• test_adopt_project_is_the_source.py:
  • test_plan_refuses_a_name_that_is_not_the_sources[hyphenated-source-retyped, typo] and test_an_org_repo_source_is_the_repository_it_names now expect the adopt-project.py plan takes the organisation from the source's origin URL and runs it through checked_value("--org", ...), so an origin whose organisation part is not ASCII or holds a space (/x/café/repo.git, /x/my repos/repo.git, caf%C3%A9) ends plan in REFUSED unsafe-value: --org is 'café', a flag nobody passed, with a remediation that explains a leading - and never says to pass --org #427 naming. They pass no --org.
  • test_a_name_read_from_a_path_says_so and test_a_name_read_from_a_path_is_renamed_at_that_path are removed. Their behaviour is reversed, and the new file covers it (test_a_path_origin_is_not_compared_and_plan_says_so, test_a_path_origin_gives_no_rename_first_remedy).
• test_adopt_git_reads.py:
  • NO_ORG is the new sentence (test_plan_without_org_refuses_a_source_that_names_none, 3 rows).
  • The ASCII control of the not-UTF-8 pair is a url (git@example.invalid:cafe/repo.git), with a not-UTF-8 url beside it. Its old spelling, ../cafe/repo.git, is a path and now names nothing. These two rows pass on main by design.
• test_adopt_split_branch.py: the GitHub stand-in is a url origin, https://github.com/testorg/Northwind.git, fetched from the bare by path into refs/remotes/origin/*. The bare's own path as origin now names nothing. These 4 tests pass on main and on this branch.
• test_shape_doctor.py: in ORIGIN_ORGS, [trailing-slash] now expects opensoft. It gains [posix-path] (/srv/git/opensoft/Atlas.git) and [file-url], both expecting None. The parity test and the comment name the one shared reading.
• test_adopt_plan_yaml_quoting.py, test_adopt_tracking_branch.py: comments only.

Revert table. Each change was reverted alone in a scratch clone of b3cc111, and the new and changed adopt files ran with the doctor's two origin tests (test_the_adoption_line_names_an_org_only_where_plan_reads_none, test_an_ancestors_origin_names_no_org_for_a_loose_directory):

reverted alone	fails
the whole change (adopt-project.py, scripts/repo_shape.py and shape-doctor.py at main 64fc9a8)	37: all 28 new; 6 changed ([hyphenated-source-retyped], [typo], test_an_org_repo_source_is_the_repository_it_names, test_plan_without_org_refuses_a_source_that_names_none[not-utf8, no-origin, no-pair]); the doctor's [trailing-slash], [posix-path], [file-url]
url: the trailing / kept	10: test_a_url_names_its_repository_however_it_ends (7), test_a_url_ending_in_a_slash_is_the_repository_it_names (2), the doctor's [trailing-slash]
path: read as its last two directories again	15: test_a_path_names_no_repository_however_it_is_written (7), ..._is_not_compared_and_plan_says_so (2), ..._reads_alike_in_either_separator[slashes], ..._gives_no_rename_first_remedy, test_a_path_origin_names_no_organisation (2), the doctor's [posix-path] and [file-url]
no NOTE for a path origin	4: ..._is_not_compared_and_plan_says_so (2), ..._reads_alike_in_either_separator (2)
#427: --org is ... again	4: test_an_org_read_from_origin_is_refused_by_where_it_came_from
#427 in adopt-not-the-source: --org X and ... again	4: test_a_mistyped_project_names_the_org_nobody_passed_by_its_source, [hyphenated-source-retyped], [typo], test_an_org_repo_source_is_the_repository_it_names
adopt-no-org: the old sentence	5: test_a_path_origin_names_no_organisation (2), test_plan_without_org_refuses_a_source_that_names_none (3)
file:// read as a url	2: ..._however_it_is_written[file-url], the doctor's [file-url]
a bracketed scp host not read	1: ..._however_it_ends[bracketed-host]
--source named as the origin	1: test_an_org_repo_source_is_the_repository_it_names
shape-doctor.py alone at main (its own copy of the old regex)	3: the doctor's [trailing-slash], [posix-path] and [file-url]

Full runs at b3cc111 (dirty 0) covered 75 test files: the 72 that grep -l -E 'adopt-project|adopt_project|\bADOPT\b' tests/*.py finds, plus test_update_shape_kept_helper.py, test_git_helper_identity.py and test_family_member_name.py. The 72 include test_shape_doctor.py, test_doctor_next_is_a_command.py, test_repo_hygiene.py and test_scaffold_leg_path_collisions.py. They ran with /usr/bin/python3 -m pytest, a stub gh first on PATH, GH_TOKEN bogus and an empty GH_CONFIG_DIR:

• with git-filter-repo on PATH: 5843 passed, 13 skipped, exit 0 (49 min 36 s)

• with it hidden (PATH=<stub>:/usr/local/bin:/usr/bin:/bin): 5530 passed, 326 skipped, exit 0 (41 min 18 s)

• The branch merges main 64fc9a8 (merge commit 2cfbe0e, clean). That includes shape-doctor.py: a git repository is told of both ways in, the adoption and the scaffold, and the adoption line ends --org <your-org> where its origin names no organisation plan can read, so pasting it is no longer refused adopt-no-org (#545) #553 (b4c1f50), which this PR now changes, so the full runs above are of the merge with that main. Main has since moved to a2243f1: openRepoShape --doctor says it writes nothing into the repository but a --placement-plan file you name, keeps its clone for the NEXT lines that name it, shows its synopsis description again, and refuses a second --placement-plan or a second path (#546) #580 and shape-doctor.py: a validator's verdict is a printed FINDING or REFUSED and any other non-zero exit is a crash, so a gate killed by a signal, ended by os._exit(1), holding conflict markers or failing argparse no longer "reports a finding", a crash of the standard's own naming script is CANNOT ANSWER instead of INVALID (naming), and a pass that printed WARNING or note lines is a note quoting them while a red row quotes its finding and never a warning (#410, #448) #557 changed shape-doctor.py and tests/test_shape_doctor.py. git merge-tree of b3cc111 with a2243f1 is clean (tree 3edcd3c5), and neither change touches the repo_shape import, origin_org, check_the_way_in or ORIGIN_ORGS. On a committed scratch merge with that tree, test_shape_doctor.py, test_doctor_next_is_a_command.py and test_adopt_origin_repository.py gave: with git-filter-repo on PATH, 456 passed, 3 skipped, exit 0; with it hidden, 452 passed, 7 skipped, exit 0.

• The doctor's line, end to end at b3cc111, with git's global config isolated:

  • For origin /srv/git/acme/Atlas.git the NEXT ends --org <your-org>. Pasted with acme, it plans with exit 0, the NOTE and org: acme.
  • For https://github.com/acme/Atlas/ it carries no --org and plans with org: acme.

Windows

• Every new row is a string git keeps as written, so the backslash rows (C:\src\acme\Atlas.git, \\server\share\...) run on every platform. Paths a test creates are tmp_path joined with clones, x/café, x/my repos and <name>.git. All are legal on NTFS, and on Windows they are drive paths through as_posix().
• The non-ASCII assertions rely on PYTHONUTF8=1, which the Windows job sets.
• Nothing counts or compares "…\n" in bytes read back from a file.
• Under a newline emulation kept outside the repository (a pytest plugin that makes text-mode Path.write_text and text-mode stdin write \r\n), the new and changed files ran at b3cc111: test_adopt_origin_repository.py, test_adopt_project_is_the_source.py, test_adopt_git_reads.py, test_adopt_split_branch.py, test_adopt_tracking_branch.py, test_adopt_plan_yaml_quoting.py, test_shape_doctor.py. Result: 1013 passed, 3 skipped and 4 failed (exit 1).
  • The 4 failures are the doctor's shim tests: test_the_shim_runs_the_script_beside_it_and_fetches_nothing, test_the_shim_from_stdin_clones_the_standard_first, test_the_installed_doctors_next_lines_run_after_it_has_exited and test_the_shim_doctor_defaults_to_the_directory_you_are_in.
  • They are the emulation's limit, not this change. They fail the same way under the emulation at main a2243f1, and pass without it at b3cc111 (4 passed). refusing_network writes its gh and curl stubs with Path.write_text, so the emulation turns their #!/bin/sh\n into #!/bin/sh\r\n, which bash cannot execute. Those 4 tests are @WINDOWS_SKIP, so the Windows job never runs them.

Not in this PR

• adopt-manifest-invalid still names a filled-in --org as a flag. An organisation read from a url that checked_value accepts but the project's validate-manifest.py refuses (~x, @a, a+b, ., ..) is refused as --org is '~x', which the project's own validate-manifest.py refuses. adopt-project.py: plan names an elector it took from git config user.name by where it came from, so with no --elected-by and no git config user.name the refusal reads --elected-by (not given, so git config user.name) is '' and no longer --elected-by is '', a flag nobody passed (#529) #567 (9f80939, adopt-project.py plan: with no --elected-by and no user.name, the refusal reads --elected-by is '', naming a flag that was never passed #529) now names such values through _manifest_flag_names(given), so an org entry in FLAG_DEFAULTS would name it by where it came from. It is left for Tracking: adopt-project.py, converting an existing repository in place #457's checklist.
• check and execute print no NOTE for a path origin. Only plan does. They compare nothing with it, as plan does not.
• Following a path origin to the repository it names on disk, and to that repository's own origin.

Known limits

• A path origin is compared with nothing. A mistyped --project or --org for a clone of a local clone or mirror plans with exit 0. The NOTE names the repository execute will act on.
• A clone of a path ending <org>/<repo> (/srv/git/acme/Atlas.git) needs --org. On main, plan read it from the path. With --org it plans, with the NOTE. The doctor's adoption line (shape-doctor.py, NOT A SHAPE ROOT on a git repository: print both ways in, as AGENTS.md says, and add --org to the adopt line when there is no origin #545, shape-doctor.py: a git repository is told of both ways in, the adoption and the scaffold, and the adoption line ends --org <your-org> where its origin names no organisation plan can read, so pasting it is no longer refused adopt-no-org (#545) #553) reads origin the same way, so for such a source it ends --org <your-org>, and pasted with the organisation filled in it plans.
• An origin rewritten by url.<base>.insteadOf to a local path is read as that path. git remote get-url applies the rewrite, so a GitHub url sent to a local mirror names nothing and is not compared.
• A url's host is not asked. https://gitlab.com/acme/Atlas.git names acme/Atlas, as on main.
• check no longer reads a path origin's remote-tracking branches as GitHub's. _places_check_asks asks refs/remotes/origin/* only where the source's origin names the assembly repository, and a path names none, so a split branch a mirror fetched from GitHub is not found by check; execute asks GitHub itself.
• A url whose organisation checked_value refuses, given another --org, is a dead end. https://example.com/café/Atlas.git with --org acme is refused adopt-not-the-source, whose remediation offers --org café, which is refused unsafe-value.

Closes #521
Closes #427

🤖 Generated with Claude Code

Summary by Sourcery

Unify origin parsing across adoption and diagnostic tools so repository checks, organization inference, and remediation guidance handle URLs and local paths consistently.

Bug Fixes:

• Correct repository detection for trailing-slash URLs and prevent local filesystem paths from being mistaken for GitHub repositories.
• Align organization inference and refusal messages with whether the value came from --org, --source, or origin.
• Make shape-doctor.py use the same origin parsing behavior as adopt-project.py.

Enhancements:

• Add plan notes explaining that path-based origins are not compared and identifying the repository execute will target.

Tests:

• Add comprehensive coverage for URL, path, organization-inference, and shape-doctor origin parsing cases.

[sourcery-ai]

Sorry @brettheap, you've used your own review budget of 250,000 diff characters for the last 7 days.

You can request another review in 1 day and 8 hours by commenting @sourcery-ai review. Upgrade to get a review now.

[sourcery-ai]

Reviewer's Guide

The PR introduces a single shared origin parser for adoption and doctor workflows: URL origins now consistently resolve to their final org/repo, filesystem and file URL origins resolve to no GitHub repository, and plan/doctor diagnostics, organization inference, source comparisons, remediation, and tests are updated to match.

Sequence diagram for plan handling a local-path origin

sequenceDiagram
    participant Plan as adopt-project.py plan
    participant Git as git remote get-url origin
    participant Parser as repo_shape.url_repository
    participant Report as Plan diagnostics
    Plan->>Git: remote get-url origin
    Git-->>Plan: filesystem path
    Plan->>Parser: url_repository(path)
    Parser-->>Plan: None
    Plan->>Report: _planned_org
    Report-->>Plan: adopt-no-org if --org is absent
    Plan->>Report: _path_origin_notes
    Report-->>Plan: NOTE that --org and --project were not compared

Loading

Sequence diagram for doctor adoption-line origin handling

sequenceDiagram
    participant Doctor as shape-doctor.py
    participant Git as git remote get-url origin
    participant Parser as repo_shape.url_repository
    Doctor->>Git: remote get-url origin
    Git-->>Doctor: URL or path
    Doctor->>Parser: url_repository(origin)
    alt repository found
        Parser-->>Doctor: org/repo
        Doctor-->>Doctor: origin_org returns org
        Doctor-->>Doctor: adoption line omits --org
    else path or no repository
        Parser-->>Doctor: None
        Doctor-->>Doctor: adoption line ends --org <your-org>
    end

Loading

Flow diagram for origin classification and repository inference

flowchart TD
    Start["Read origin"] --> Path{"origin_is_a_path?"}
    Path -->|"yes: filesystem or file://"| None["No GitHub repository or organisation"]
    None --> PlanNote["plan: do not compare --org/--project; print NOTE"]
    None --> OrgPrompt["plan/doctor: require --org"]
    Path -->|"no: URL or scp-style origin"| Normalize["Strip trailing / and one .git"]
    Normalize --> Pair["Take final org/repo"]
    Pair --> Compare["plan: infer organisation and compare source"]
    Pair --> DoctorOrg["doctor: omit --org from adoption line"]

Loading

File-Level Changes

Change	Details	Files
Centralize origin parsing so URLs consistently identify their trailing repository while filesystem paths identify no GitHub repository.	Add shared helpers to classify paths, including drive, UNC, separator, and file URLs. Parse scheme, SCP, bracketed-host, credentials, ports, trailing slashes, and optional .git correctly. Make adoption source comparison and organization inference use the shared parsing result.	scripts/repo_shape.py adopt-project.py shape-doctor.py
Revise plan behavior and diagnostics for path origins and inferred organizations.	Stop comparing --org/--project with path origins and emit a NOTE describing the repository execute will target. Require --org when a path origin supplies no organization, with updated adopt-no-org messaging. Identify organizations inferred from origin or --source in unsafe-value and adopt-not-the-source errors, and update remediation text. Remove path-specific rename-first handling because paths no longer name repositories.	adopt-project.py
Keep the shape doctor’s adoption guidance behaviorally identical to plan.	Replace the doctor’s copied origin regex with repo_shape.url_repository. Make the adoption line request --org for paths and other origins from which plan cannot infer an organization. Update doctor documentation and parity coverage for trailing-slash, POSIX-path, and file-url cases.	shape-doctor.py tests/test_shape_doctor.py
Expand and realign regression coverage around origin parsing and changed adoption semantics.	Add parameterized coverage for URL forms, path forms, path notes, organization inference, unsafe values, and source naming. Update existing tests for path origins, URL trailing slashes, source-origin wording, and local fetch stand-ins. Document and validate the shared parsing contract across adoption and doctor tests.	tests/test_adopt_origin_repository.py tests/test_adopt_git_reads.py tests/test_adopt_project_is_the_source.py tests/test_adopt_split_branch.py tests/test_adopt_tracking_branch.py tests/test_adopt_plan_yaml_quoting.py tests/test_shape_doctor.py

Assessment against linked issues

Issue	Objective	Addressed	Explanation
#427	When plan infers an organisation from the source’s origin and checked_value refuses it, identify the value as inferred rather than as a user-supplied --org, and provide remediation that tells the user to pass --org while retaining the refusal and exit code.	✅
#521	Read the repository identity from a local-path origin so the source-name check can distinguish the right source from a mistyped --project.	❌	The PR deliberately treats every path origin as naming no repository. It skips the source-name comparison, so a mistyped --project can still plan successfully; the PR documents this as a known limit.
#521	Read the repository identity from a Windows backslash-path origin so the source-name check is not skipped.	❌	The PR recognizes backslash paths as paths, but deliberately returns no repository identity for them. It reports that no comparison was made rather than comparing the source name.
#521	Recognize the repository named by a URL with a trailing slash, including .git/, so a mistyped --project is refused.	✅
#521	Name an organisation read from the source when refusing unsafe values or a mismatched project, and direct the user to pass --org.	✅

---

Tips and commands

Interacting with Sourcery

• Trigger a new review: Comment @sourcery-ai review on the pull request.
• Continue discussions: Reply directly to Sourcery's review comments.
• Generate a GitHub issue from a review comment: Ask Sourcery to create an
  issue from a review comment by replying to it. You can also reply to a
  review comment with @sourcery-ai issue to create an issue from it.
• Generate a pull request title: Write @sourcery-ai anywhere in the pull
  request title to generate a title at any time. You can also comment
  @sourcery-ai title on the pull request to (re-)generate the title at any time.
• Generate a pull request summary: Write @sourcery-ai summary anywhere in
  the pull request body to generate a PR summary at any time exactly where you
  want it. You can also comment @sourcery-ai summary on the pull request to
  (re-)generate the summary at any time.
• Generate reviewer's guide: Comment @sourcery-ai guide on the pull
  request to (re-)generate the reviewer's guide at any time.
• Resolve all Sourcery comments: Comment @sourcery-ai resolve on the
  pull request to resolve all Sourcery comments. Useful if you've already
  addressed all the comments and don't want to see them anymore.
• Dismiss all Sourcery reviews: Comment @sourcery-ai dismiss on the pull
  request to dismiss all existing Sourcery reviews. Especially useful if you
  want to start fresh with a new review - don't forget to comment
  @sourcery-ai review to trigger a new review!

Customizing Your Experience

Access your dashboard to:

• Enable or disable review features such as the Sourcery-generated pull request
  summary, the reviewer's guide, and others.
• Change the review language.
• Add, remove or edit custom review instructions.
• Adjust other review settings.

Getting Help

• Contact our support team for questions or feedback.
• Visit our documentation for detailed guides and information.
• Keep in touch with the Sourcery team by following us on X/Twitter, LinkedIn or GitHub.

[brettheap]

@codex review

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	⚠️ Failed 2026-10-15T14:33:01.326543Z	b3cc111	Manual request

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.
To continue using code reviews, you can upgrade your account or add credits to your account and enable them for code reviews in your settings.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud