Skip to content
Skills Library
Security · Learning & explanation

secret-entry

/secret-entry

Use when an access needs a secret the owner must type (an API key), when adding or removing an access, or when a route says that access is missing a variable. Explains how the key is entered without the chat or the model ever seeing it.

Licence
Apache-2.0
Last repository change
04.11.2026

Assessment and security

96/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content31 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
3
Scripts
0
HTML comments
0
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 04.11.2026, without references to retired models or APIs.

Licence

Apache-2.0, otwarta: można używać i zmieniać, z zachowaniem licencji i informacji o zmianach.

What it does and when to use it

Use when an access needs a secret the owner must type (an API key), when adding or removing an access, or when a route says that access is missing a variable. Explains how the key is entered without the chat or the model ever seeing it.

Original author description

Use when an access needs a secret the owner must type (an API key), when adding or removing an access, or when a route says that access is missing a variable. Explains how the key is entered without the chat or the model ever seeing it.

How it works

SKILL.md has no level-two headings · 21 lines, 427 words

SKILL.md preview
---
name: secret-entry
description: Use when an access needs a secret the owner must type (an API key), when adding or removing an access, or when a route says that access is missing a variable. Explains how the key is entered without the chat or the model ever seeing it.
---

# Secret entry (a key goes from the owner's keyboard to the node's credential store)

The owner types a key into a **native hidden-field window of the operating system** on the computer he is sitting at. It never goes through a chat, a tool argument, a card or a model. The coordinator reads it from that window, seals it into one write task to the node over the paired channel, and the node keeps it in **its own credential store** (Windows Credential Manager, macOS Keychain, Linux Secret Service). There is no other place: with no real store the entry is refused with `credential_store_missing`, and nothing is written to a file.

What you can do, and what you will see:

| Goal | Tool | What comes back |
|---|---|---|
| Add an access (a key) on a machine | `mas_add_access` | the plan as text; the harness itself opens the approval dialog (Touch ID) on the owner's screen and, right after he approves, the hidden-field window for the key. You call nothing; follow with `mas_job`. If no dialog could be opened, `mas_job` gives the `intent_id` for `mas_request_approval` |
| Remove an access | `mas_remove_access` | the plan (what is set aside, how to undo it); the owner approves in the system dialog; with `undo: true` it brings the key back |
| See whether a route is missing a key | `mas_explain_route` | `access missing: NAME` for the machine; the NAME of the variable, never a value |

Rules you follow:
- You never ask the owner to paste a key into the chat and never accept, repeat or store one. If he pastes one, tell him to revoke it at its provider (it is in the chat now) and to enter a new one in the window.
- You never see a value. The only thing you learn is "access present" or "access missing: NAME".
- Removing an access moves the key aside inside the store, so it can be undone without the key passing through anyone. Approve and undo both need the owner's own gesture in the system dialog; a tool call cannot give it.
- A machine whose store is missing (for example a Linux computer with no Secret Service running) cannot hold a key: say so and stop; do not suggest a file or an environment variable as a workaround for a key the harness should hold.

`references/prompts.json` has the window for each operating system as data. A new operating system or window program is a new row there, with no code change.
Files in this package (3)
  • SKILL.md 3 KB
  • skill.meta.json 1 KB
  • references/prompts.json 3 KB

Versions

The database has no other skills named secret-entry.

Database snapshot: 13.11.2026 · package version f141a59b7db7 (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 100leak-response
    Step-by-step response when a secret, token or private key was committed or published — rotate first, then check for abuse, then clean history. Use the moment a live secret is found, or when the user says "I pushed my API key".
  • 100secret-hygiene
    How to handle an API key, token, password or connection string in a repository - where the value goes, what gets committed, and what to do when one has already been committed or pushed. Use when wiring up authentication or a third-party API, when adding config, when a write was…
  • 100rotate-secret
    Replace a secret (API key, OAuth credential, generated token) safely in an existing project. Lists the rotatable secrets, points the user to the right provider dashboard with click-by-click instructions, captures the new value (or auto-generates it for self-managed secrets), and…
  • 100setup-secret
    Secret access wiring and manifest authoring. Use when a workload needs to read a secret, the user asks to create a secret or generate secret YAML, configure a pull secret, or fix a deployment paused on a secret reference.
  • 100leaked-secret
    Respond to a leaked API key, token, password or private key in a commit, log, screenshot, issue or chat by rotating it first, then removing it, checking for use and closing the gap. Use it the moment a secret turns up where it should not, even in a commit that was never pushed.
  • 90secret-leak-response
    Respond to a leaked secret (API key, cloud key, token, password, private key, connection string) found in a public repo, logs, chat, paste site or container image — validate, revoke/rotate at the source, check for misuse in provider logs, purge copies, and prevent recurrence.…

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.