- Home
- Skills
- APIs & Backend
- Storage Sanitization Policy, Method Catalog and Receipts
Storage Sanitization Policy, Method Catalog and Receipts
A product that lets people "shred" files will be asked for a certificate, and the honest answer depends on media, copies, keys and authority, not on how many passes were written.
$29
Works with the AI Agents you already use
Storage Sanitization Policy, Method Catalog and Receipts
Example session with this skill installed
A customer wants to shred a contract PDF stored on our SSD-backed tenant volume and get a certificate. The tenant admin approved it; there is a nightly backup that still has the file. What can we honestly claim, and generate the receipt.
- Read your context and instructions
- Compiled the storage sanitization policy
Summary
The policy denied the overwrite "clear" claim and allowed only a logical deletion with a receipt that says exactly that.
Decision
evaluateShredRequest: media is SSD and copy-on-write is not proven absent, so an overwrite pass cannot be claimed as clear; the nightly backup is unresolved, so no purge-level claim is possible either.- Allowed outcome:
logicaldeletion with assurancenone-beyond-logical, confirmation phrase "Delete the file; copies may remain in backups". - Required evidence keys for the receipt: object digest, tenant authorization id, backup disposition note.
Receipt
sanitization-receipt/v1, Ed25519-signed over canonical JSON. The claim sentence printed on it: "The object was deleted from the live system. No assurance is made about backups, snapshots or the underlying media." Verification with the public key: valid.
Tests
node --test scripts/sanitization.test.mjs: 8 passed.
Connects securely to your tools. The creator never sees your data.
About this skill
A product that lets people "shred" files will be asked for a certificate, and the honest answer depends on media, copies, keys and authority, not on how many passes were written. This skill is the decision layer between a delete button and a claim.
Catalog (sanitization-methods.mjs): 17 profiles with family (logical, overwrite, purge, destroy), status (current guidance, historical, operator-defined, controlled service), executable flag, compatible media, the exact pass plan and the claim sentence to print. Verified single-pass with read-back and ITSP random 3-pass as current guidance; HMG IS5, DoD-style 3 and 7, Schneier, VSITR, RCMP and the exact Gutmann 35-pass sequence as visibly legacy and magnetic-only; custom random 1 to 99; cryptographic erase, device sanitize commands, degaussing, disintegration and incineration as controlled methods the client may record but never self-certify.
Policy (sanitization-policy.mjs): evaluateShredRequest fails closed on legal hold, unexpired retention, missing tenant authorization or user presence, and unresolved backups; denies overwrite clear on SSD, flash, virtual, unknown or remapped storage and wherever copy-on-write or snapshots are not proven absent; requires a unique destroyable key and known key-backup disposition for crypto erase, device-owner authority for managed wipes, and a destruction operator with evidence for physical attestation; returns the assurance level and the evidence keys the receipt must carry. Exact confirmation phrases.
Receipts (sanitization-receipt.mjs): a receipt is created only for a completed, allowed operation whose assurance matches the decision and whose evidence covers every required key; identities are SHA-256 digests; Ed25519 signature over canonical JSON; verification for anyone with the public key.
Reference: the method catalog with the standards each profile rests on (NIST SP 800-88 Rev. 2, IEEE 2883, UK NCSC, Canadian ITSP.40.006, NSA/CSS PM 9-12) and the non-negotiable claim boundary; a threat model with the control for each threat.
8 offline tests. Node 20+, no packages, no network. No adapters included: it decides and records, it does not erase.
Does not perform I/O or deletion. Requires custom adapters for hardware execution. No network access. Compliance depends on the accuracy of evidence provided by your implementation.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 19 days ago
Requires
Permissions
Shell to run the node scripts and tests; read/write files for policy input, receipt output and the Ed25519 key files. No network access, no environment variables.