Apply a five-point security checklist to external outputs before constructing Bash commands.
npx skills add https://github.com/aldengolab/lorist --skill sec-unit
Please help me install this Agent Skill. Skill: sec-unit Source: https://github.com/aldengolab/lorist/tree/main/skills/sec-unit Command: npx skills add https://github.com/aldengolab/lorist --skill sec-unit
A five-point security checklist helps ensure you don't construct dangerous Bash commands when results come from kubectl, argocd, helm, WebFetch, or git log. It prevents prompt-injection, data leakage, and misconstructed pipelines by requiring explicit authorization before actions.
Ask sec-unit to run the five-point security checklist immediately after receiving external output and before composing any Bash command.
High-intent search queries and answers about installing and using this skill.
Preventing prompt injection from kubectl output requires screening external results with a five-point security checklist before any Bash command is built. This process detects instruction-in-data positions and halts execution on flagged conditions, requiring explicit user authorization to proceed.
Data exfiltration risk in multi-step pipelines occurs when external outputs from tools like helm or argocd contain malicious URL-path substitutions that covertly transmit sensitive data. Detecting this risk involves checking outputs against a security checklist before authorizing downstream command execution.
Securing git log output involves running the external output through a five-point security checklist before constructing Bash commands. This identifies potential data injections and inline execution risks, halting the pipeline to require explicit authorization before any action proceeds.
Claude Code works with a security gating layer by applying a checklist to WebFetch outputs before Bash command construction. This layer prevents misconstructed pipelines and data leakage by requiring explicit user authorization before proceeding with any flagged automated infrastructure tasks.
You should halt command execution to prevent unsafe Bash command construction whenever a five-point security checklist flags conditions like instruction-in-data positions, URL-path exfiltration risks, or inline executions in external outputs. The process requires explicit user authorization before proceeding.
phuetzConsolidate kubectl, Helm, and ArgoCD workflows for Kubernetes cluster administration.
mixomatDelete local git branches marked [gone] while skipping worktree branches.
bomberoxenviosdosruedasAutomate DevOps workflows with Terraform, GitHub Actions, Kubernetes, and Prometheus.
kwoekelAnalyze Claude skills for security risks and vulnerabilities with Python scripts.
