node-certify
Skill4 starsChanged 4 months agoMITView on GitHub
Use when fresh certificates are requested for one or more Officina nodes. Do not use merely to check certificate currentness or canonical node hashes.
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-driftSkill
- notation-reviewSkill
- online-calendarSkill
- pdf-to-markdownSkill
- prepare-handoffSkill
- proof-auditSkill
- recurring-tasksSkill
- refactor-nodeSkill
- regenerate-blueprintsSkill
- relocate-nodesSkill
- 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
4 rules. About 0 are common, 2 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 63
Write each exact final JSON result to its own report file and call complete PREFIX TASK_ID --report-file FILE, then claim again to refill …
- Very specific: line 69
Only after scheduler status is complete, invoke the declared mechanical certify interface for the requested target and exact reviewed …
- Distinctive: line 86
Never ask it to sign caller-supplied certificate data.
- Distinctive: line 93
If synchronization, validation, semantic audit, hashing, signing, or post-write verification fails, retain earlier valid append-only …
≈ “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.
node-certify._rtx.interface.certify— Certify exact v6 module closures by skipping current nodes and appending signed certificate histories for stale nodes at an explicit reviewed repository commit.- Caller:
node-certify - Version: 2
- Security level: 2
- Alternative:
defaultArguments JSON (replace labels with actual values). Omit optional positionals and options that are not needed. {"options": {"--allow-non-atomic": true, "--json": true, "--reviewed-commit": "COMMIT", "--reviewed-repository": "ROOT"}, "positionals": ["certify", "target..."], "stdin": null} Required options: []; positional arity: 1..unbounded; stdin: forbidden
- Caller:
node-certify._rtx.interface.semantic-audit-scheduler— Initialize or transition one locked semantic-audit run.- Caller:
node-certify - Version: 1
- Security level: 2
- Alternative:
defaultArguments JSON (replace labels with actual values). Omit optional positionals and options that are not needed. {"options": {}, "positionals": ["OPERATION", "PREFIX", "TASK_ID", "options"], "stdin": null} Required options: []; positional arity: 2..unbounded; stdin: forbidden
- Caller:
node-drift._rtx.interface.drift-status— Read signed certificate currentness, exact structured drift causes, and the dependency-first stale worklist for exact or installed v6 modules without writing certification state.- Caller:
node-certify - Version: 4
- Security level: 2
- Alternative:
defaultArguments JSON (replace labels with actual values). Omit optional positionals and options that are not needed. {"options": {"--all": true, "--dag-file": "PATH", "--json": true, "--repo-root": "ROOT", "--skill-root": "ROOT"}, "positionals": ["target..."], "stdin": null} Required options: []; positional arity: 0..unbounded; stdin: forbidden
- Caller:
Instruction Interfaces:
These are LLM-readable instruction surfaces. Read and follow them directly; do not invoke the MCP server for them.
node-certify.source.audit-behavioral-source.interface.audit@2— Audit one behavioral source and return bounded semantic evidence and a verdict.node-certify.source.audit-interface.interface.audit@2— Audit one source interface and return bounded semantic evidence and a verdict.node-certify.source.audit-module.interface.audit@2— Audit one module and return bounded semantic evidence and a verdict.
<!-- END BLUEPRINT INTERFACES -->Certification algorithm
Resolve the requested target and hold its reviewed repository and commit stable. Then:
- Invoke
node-drift._rtx.interface.drift-status@4once in JSON mode with--dag-file, and save its JSON result for scheduler initialization. - Initialize
node-certify._rtx.interface.semantic-audit-scheduler@1from the DAG and drift result. The scheduler, not the LLM, owns dependency traversal and audit readiness. - Determine available subagent slots
K, excluding the orchestrator; use one if unknown. Callclaim --capacity K. For each returned item, spawn one fresh subagent and passinput_fileunchanged. Mapinterfacetonode-certify.source.audit-interface.interface.audit@2,behavioral-sourcetonode-certify.source.audit-behavioral-source.interface.audit@2, andmoduletonode-certify.source.audit-module.interface.audit@2. Never reuse a subagent for another task. - Write each exact final JSON result to its own report file and call
complete PREFIX TASK_ID --report-file FILE, then claim again to refill the pool. If no task is returned while work remains in progress, wait. On spawn failure or worker loss callfail. Stop all remaining work on malformed output,reject,abort, or scheduler failure; do not infer readiness or recursively audit dependencies. - Only after scheduler status is
complete, invoke the declared mechanicalcertifyinterface for the requested target and exact reviewed repository and commit. It independently recomputes currentness, skips current nodes, route-smokes the stale worklist, and issues stale nodes dependency-first.
With a matching basis, reuse authenticated semantic evidence when its facet
local hash or module node hash, input manifest, ordinary dependencies, and
governing semantic certified-under claim still match. The mechanical certify
claim remains required for currentness and issuance but does not by itself
require semantic re-audit. A remainder-facet cause belongs to
audit-behavioral-source; it does not create a remainder interface.
Schema validity is necessary but does not establish semantic accuracy. The audit interfaces own semantic judgment. The mechanical interface invokes the repository validator runner, reconstructs every payload field, computes hashes, signs, appends certificate history, and performs post-write drift verification. Never ask it to sign caller-supplied certificate data.
Existing logs must be canonical, schema-valid, signature-valid, unbroken, and a dependency-first prefix of the exact closure. New certificates require tracked inputs to match the reviewed commit and included local inputs to remain byte-stable. Dirty or unready state may be reported but must not be certified.
If synchronization, validation, semantic audit, hashing, signing, or post-write verification fails, retain earlier valid append-only history, report the exact failure, and do not claim current certification.
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.

