skills.rest
Noetfield-Systems
Noetfield-SystemsOFFICIAL
Explore all skills by this author →
WebsiteGitHub

skill-006-ask-before-implement

Enforces founder confirmation before any disk edit in SourceA ecosystem agent sessions.

Updated Jun 20, 2026
Search Tags:#approval gate#agent governance#change control#session protocol#confirmation workflow#disk edit#filing registry
One-click install
npx skills add https://github.com/Noetfield-Systems/SourceA --skill skill-006-ask-before-implement-noetfield-systems
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: skill-006-ask-before-implement
Source: https://github.com/Noetfield-Systems/SourceA/tree/main/.cursor/skills/skill-006-ask-before-implement
Command: npx skills add https://github.com/Noetfield-Systems/SourceA --skill skill-006-ask-before-implement-noetfield-systems

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? AI agents in the SourceA ecosystem can make unauthorized disk edits without explicit founder approval, causing incidents like INCIDENT-010 and INCIDENT-011. This Skill enforces a mandatory ask-before-implement gate so no file is created or modified without an explicit order. ## Core Features & Use Cases - Pre-edit confirmation gate: Requires the agent to restate its understanding, propose options A/B/C, and wait for an explicit order (SAVE TO:, EDIT ALLOWED:, ASF:, WORK:) before any disk mutation. - Filing registry resolution: Resolves pathless SAVE/LOCK/FILE intents through the agent filing registry script instead of asking the founder for exact paths. - Session-end protocol: Mandates a done/open/blocked summary plus a "What is your next move?" question at the end of every session. - Use Case: A founder says "save this" without a path; the agent runs the filing registry gate, resolves the route, and only writes the file after the registry returns a valid path and next steps. ## Quick Start Activate this skill at the start of every SourceA ecosystem agent session so the agent asks for confirmation and suggested options before making any disk edit.

Frequently Asked Questions about skill-006-ask-before-implement

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I prevent an AI agent from editing files without permission?▼

Use an ask-before-implement gate that blocks all disk mutations until the user gives an explicit order. The agent must restate its understanding, propose options A/B/C, and wait for commands like SAVE TO: or EDIT ALLOWED: before writing.

How to handle pathless save commands from a user in an agent workflow?▼

Run the agent filing registry gate script with the user's intent message to resolve a route_id, target path, and next steps. Only ask the user for clarification when the registry returns no route or ambiguous routes.

What commands authorize disk edits in this approval workflow?▼

Authorized triggers are SAVE TO: or SAVE AS: with a path, EDIT ALLOWED: plus ACTION:, ASF: with an explicit phrase, WORK: with sa-* or INBOX scope, registry-resolved filing intents, or a yes/option letter after an ASK.

What happens when rules conflict in this agent governance skill?▼

On rule conflict, the agent must run the skill-007-ecosystem-conflict-resolution skill first before proceeding. This ensures conflicting governance rules are resolved through a dedicated escalation path rather than ad-hoc decisions.

When does this approval gate not apply to agent edits?▼

Bounded exceptions exist: WORK: with sa-* or INBOX scope allows normal-speed builds within scope, and SAVE TO: with an explicit path permits exactly one new file at that path. All other cases require explicit founder orders.

Related Skills

Explore More Software Engineering Skills→
nicepkgnicepkg
276

capture-triage

Classifies Drafts Pro inbox captures and routes them to daily Ready section.

Official
Advanced
MythologIQMythologIQ

living-playbook-sre

Map incident categories to operational runbooks with owners and rollback controls.

Community
Intermediate
mst-software-vnmst-software-vn
1

using-superpowers

Establishes rules for discovering and invoking skills before any response.

Official
Intermediate
antonyfmunozantonyfmunoz

adversarial-review

Run adversarial code review on changed source files and verify with py_compile.

Community
Advanced