Apply archive

Skill aisflows/skill-cleaner/skills/apply-archive

Clean messy AI-agent skill folders into a usable SKILL.md library.

Install

npx -y skills add aisflows/skill-cleaner --skill apply-archive

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Use as step 5 of AIS FLOWS Skill Library Cleaner to create the clean-library output from an approved plan, with dry-run, backup, archive, no permanent delete, exact operations, and APPLY_REPORT outputs.

SKILL.md

2.8 KB, 588 tokens by cl100k_base, as published. Nobody here has run it

Skill Library Cleaner / Apply & Archive

Role

Apply the approved cleaner plan to an output folder.

Input:

machine/CANONICAL_LIBRARY_PLAN.json
approved output root
operation allowlist

Output:

reports/APPLY_PLAN.md
machine/APPLY_PLAN.json
reports/APPLY_REPORT.md
machine/APPLY_REPORT.json
clean-library/
backup/

Hard Gate

Always do dry-run first.

Do not write until there is explicit approval for:

source root;
output root;
operation types;
affected paths.

machine/APPLY_PLAN.json must include:

{
  "approval_state": "dry-run-only | approved-result-folder-only | approved-specific-operations",
  "approved_by": "user | blocked | unknown",
  "approved_scope": "exact output root and operation allowlist"
}

If approval_state is dry-run-only, blocked, missing, or ambiguous, stop after APPLY_PLAN and do not apply library writes.

In dry-run mode, reports and machine JSON may be written only to the approved result folder. Do not create or modify clean-library, backup, blocked, archive, or any source-folder content unless the approval state allows that exact write scope.

Before any write-run, compare the write request against the dry-run plan:

source root must match;
output root must match;
operation types must match the allowlist;
affected paths must match;
approved_by must be user;
permanent delete count must be 0;
backup must be prepared before writes.

If any check fails, stop and report REWORK or FAIL.

Allowed Operations

copy to canonical;
copy to references;
copy duplicate to archive;
copy risky item to blocked;
write backlog note;
write approved normalized SKILL.md;
create final reports.

Permanent delete is not allowed in Skill Cleaner v0.1.0.

Prefer copy over move. Moving source items is only allowed on a copied sandbox or with explicit approved-specific-operations approval.

Required Operation Fields

Each operation must include:

operation id;
type;
source path;
target path;
reason;
risk;
rollback note;
approval status.
before/after path state when the operation changes files.

Operation approval must be explicit per operation. Do not infer approval from the existence of a plan.

machine/APPLY_REPORT.json must include:

write_step_executed;
source_mutation_detected;
permanent_delete_count;
operations_applied[];
operations_skipped[];
rollback_available.

Pass Condition

The output folder exists, duplicates are off the active shelf, risky items are blocked, and every write has an operation record.