relocate-nodes
Skill4 starsChanged 4 months agoMITView on GitHub
Use when registered Officina nodes or their owned files must be moved while mechanically updating blueprint ownership, references, generated artifacts, and callers. Do not use for behavioral refactoring or certification.
In this repository45 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- bib-auditSkill
- bootstrap-dispatcher-runtimeSkill
- ci-debugSkill
- cloud-filesSkill
- connect-googleSkill
- daily-planSkill
- distill-to-ruttersSkill
- email-clientSkill
- email-triageSkill
- find-handoff-candidatesSkill
- formal-prose-reviewSkill
- get-weatherSkill
- git-workflowSkill
- hook-makerSkill
- initialize-tddSkill
- install-launchersSkill
- latex-workshopSkill
- list-managerSkill
- llm-wakeupSkill
- loose-modeSkill
- make-tex-docstringSkill
- math-dependency-graphSkill
- milestone-loggingSkill
- node-certifySkill
- node-driftSkill
- notation-reviewSkill
- online-calendarSkill
- pdf-to-markdownSkill
- prepare-handoffSkill
- proof-auditSkill
- recurring-tasksSkill
- refactor-nodeSkill
- regenerate-blueprintsSkill
- semantic-integrationSkill
- send-feedbackSkill
- setup-interface-managerSkill
- skill-makerSkill
- technical-flow-reviewSkill
- tight-modeSkill
- tool-applicabilitySkill
- update-standardsSkill
- using-compassSkill
- wrap-upSkill
The file
We split this skill into the rules it gives an agent
10 rules. About 0 are common, 6 are distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈6
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 54
Add scoped python_modules only when the Python import mapping cannot be proved from repository configuration and parsed imports.
- Distinctive: line 59
Leave apply disabled. This report is the exhaustive mechanical replacement recipe.
- Distinctive: line 70
Ask the user for help removing false positives: Before asking for decisions, render every review unit with its path, section when present, …
- Distinctive: line 77
Never infer persisted-state, compatibility, or behavioral policy from an address change.
- Very specific: line 79
Encode the answer compactly: Prefer one default_disposition plus exact path-level disposition_overrides.
- Very specific: line 85
Put relevant changes that contain no reported old identifier in supplemental_edits.
- Distinctive: line 95
Never apply a mechanical subset before user review.
- Very specific: line 100
Run the identical manifest as a target-side postflight without --apply.
- Distinctive: line 104
Use the report's machine timings when discussing performance.
- Distinctive: line 109
Never invoke certification or installation.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
<!-- BEGIN BLUEPRINT INTERFACES -->Generated from
blueprint.yaml. Do not edit this block by hand.
Executable Interfaces:
Send the required caller (caller skill), interface, version, and arguments; optional dry_run defaults to false. Compact uses ordered positionals plus an option mapping; ordered raw argv uses positionals: [] plus every argv token in list options. Never mix forms.
relocate-nodes._rtx.interface.build-review-packet— Group one exhaustive relocation preflight for user-assisted false-positive filtering.- Caller:
relocate-nodes - Version: 1
- Security level: 2
- Alternative:
defaultArguments JSON (replace labels with actual values). Omit optional positionals and options that are not needed. {"options": {"--output": "OUTPUT", "--report": "REPORT", "--root": "ROOT"}, "positionals": [], "stdin": null} Required options: ["--output", "--report", "--root"]; positional arity: 0..0; stdin: forbidden
- Caller:
relocate-nodes._rtx.interface.relocate— Preflight or publish one manifest-driven registered-node relocation as one recovery-backed failure-atomic change set.- Caller:
relocate-nodes - Version: 2
- Security level: 2
- Alternative:
defaultArguments JSON (replace labels with actual values). Omit optional positionals and options that are not needed. {"options": {"--apply": true, "--manifest": "MANIFEST", "--report": "REPORT", "--root": "ROOT"}, "positionals": [], "stdin": null} Required options: ["--manifest"]; positional arity: 0..0; stdin: forbidden
- Caller:
<!-- END BLUEPRINT INTERFACES -->Machine routes
- Review packet:
relocate-nodes._rtx.interface.build-review-packet@1 - Preflight and atomic apply:
relocate-nodes._rtx.interface.relocate@2
Workflow
The relocation has three stages: list the mechanical recipe, let the user remove false positives and refine relevant prose, then publish the combined changes through one apply invocation. Stages 1 and 3 are mechanical; do not turn them into review loops.
List the mechanical recipe. Build one schema-v3 manifest whose
relocationsentries are complete physical moves. A nested relocation is one entry, never one move per path segment:schema_version: 3 relocations: - from: skills/a/b/c to: skills/a/d/eAdd scoped
python_modulesonly when the Python import mapping cannot be proved from repository configuration and parsed imports.- Invoke the injected relocation interface for a read-only preflight with the exact repository root, schema-v3 manifest, and a report path outside the selected repository. Leave apply disabled. This report is the exhaustive mechanical replacement recipe; do not mutate the repository during Stages 1 or 2.
Build the user review packet. Invoke the injected review-packet interface with the absolute repository root and external report/output paths.
The interface writes the full JSON packet and prints its complete compact human rendering. Present that rendering directly; do not spend an LLM turn reconstructing it. The packet groups Markdown by heading and other text by file, retaining every occurrence and its suggested decision. Historical plans and specifications are suggested preserves; everything else is suggested rewrite.
- Ask the user for help removing false positives. Before asking for decisions, render every review unit with its path, section when present, suggested decision, and occurrence count. Never substitute aggregate counts or an explanation for the packet itself. Show individual occurrence contexts when a unit is ambiguous or the user asks. Then ask about whole categories, files, or sections. The default is the mechanical rewrite. The user may preserve an irrelevant unit or refine awkward prose. Never infer persisted-state, compatibility, or behavioral policy from an address change.
Encode the answer compactly. Prefer one
default_dispositionplus exact path-leveldisposition_overrides; the machine expands these rules into the occurrence ledger. Usesemantic_decisionsonly for individual exceptions or a user-supplied exact replacement. Precedence is individual decision, path override, then default.Put relevant changes that contain no reported old identifier in
supplemental_edits. Each entry names the projected targetpath, one nonempty exactexpectedstring, and itsreplacement. The precondition must identify exactly one old or already-replaced string; ambiguous or missing edits fail preflight. Supplemental edits join the same transaction.- Rerun preflight with the reviewed manifest. Continue only when
unaccounted_semantic_occurrencesis empty and every planned write and delete is intended. - Apply once. Supply the combined mechanical recipe and all user-assisted decisions to the same injected relocation interface with apply enabled. Never apply a mechanical subset before user review. Publication is one recovery-backed failure-atomic transaction: it locks the repository, revalidates the preflight baseline, records external rollback state, and restores every touched path if publication or target-side postflight fails. An interrupted marker is recovered before the next invocation plans work.
- Run the identical manifest as a target-side postflight without
--apply. Require no planned writes or deletes, no unaccounted occurrences, and no errors.
Use the report's machine timings when discussing performance. An apply report separates initial planning, transactional writes, blueprint-graph verification, target-side postflight, and total time; do not attribute later test time to ledger application.
Never invoke certification or installation. Completed manifests are temporary inputs and are not retained as repository history.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

