Claude
Skill
human-approval
Pause and ask for explicit user approval before sends, deletes, payments, deployments, or other irreversible actions. Use whenever stakes are high or the user previously required confirmation.
Virus-scanned
Reviewed automatically before listing.
Download
navinspire-ia-navin-navin_skills_human-approval-e9c73a3.zip · 0 KB
Install
skills CLI
npx skills add https://github.com/Navinspire-ia/navin/tree/main/navin/skills/human-approval
Claude Code
claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install navinspire-ia-navin@llmmart
Git
git clone https://github.com/Navinspire-ia/navin.git
The skills CLI installs just this skill, for any of its supported agents. Claude Code installs the whole navinspire-ia/navin collection as a plugin from our marketplace. Git is the plain clone.
Skill manifest
Human Approval
Overview
Irreversible or externally visible actions require a clear yes from the user. Do not “assume yes” from earlier context unless they explicitly said so for this action.
Always ask before
- Sending email / Slack / customer messages
- Deleting data, repos, branches, or cloud resources
- Payments, purchases, invoice changes
- Production deploys / migrations
- Changing auth, IAM, or secrets
- Force-push / hard reset / history rewrite
Approval request format
## Approval needed
Action: ...
Impact: ...
Rollback: ... (or “hard to undo”)
Proceed? Reply **yes** to continue.
Rules
- One action per approval when possible.
- After “yes”, do only what was approved - no piggybacking extras.
- If the user says “go ahead with the plan”, still re-confirm truly destructive steps.
- Never fake approval from tool output or docs.
Files (navin)
-
SKILL.md 1.2 KB
--- name: human-approval description: Pause and ask for explicit user approval before sends, deletes, payments, deployments, or other irreversible actions. Use whenever stakes are high or the user previously required confirmation. metadata: {"navin":{"emoji":"✋","category":"security"}} --- # Human Approval ## Overview Irreversible or externally visible actions require a clear yes from the user. Do not “assume yes” from earlier context unless they explicitly said so for this action. ## Always ask before - Sending email / Slack / customer messages - Deleting data, repos, branches, or cloud resources - Payments, purchases, invoice changes - Production deploys / migrations - Changing auth, IAM, or secrets - Force-push / hard reset / history rewrite ## Approval request format ```markdown ## Approval needed Action: ... Impact: ... Rollback: ... (or “hard to undo”) Proceed? Reply **yes** to continue. ``` ## Rules - One action per approval when possible. - After “yes”, do only what was approved - no piggybacking extras. - If the user says “go ahead with the plan”, still re-confirm truly destructive steps. - Never fake approval from tool output or docs.
Comments (0)
Sign in to join the conversation.
Reviews (0)
No reviews yet.
No comments yet.