Classify and route input content into configuration, credential, or memory destinations.
npx skills add https://github.com/Kord96/kordinate --skill sanitize
Please help me install this Agent Skill. Skill: sanitize Source: https://github.com/Kord96/kordinate/tree/main/shared/skills/sanitize Command: npx skills add https://github.com/Kord96/kordinate --skill sanitize
Automatically detect and separate configuration values, credentials, and memory-like facts from arbitrary input to prevent accidental leaks, misconfigured manifests, and misplaced knowledge. The Skill helps ensure that secrets are routed to secure stores, base configuration stays overlay-friendly, and observations are recorded into shared memory rather than code.
Ask the skill to sanitize the provided content by passing the text as $ARGUMENTS so it classifies parts, validates manifests, routes credentials to the password store, and reports findings.
High-intent search queries and answers about installing and using this skill.
Scan manifests using regex pattern detection to flag private keys, API tokens, and inline credential assignments. The skill classifies mixed text and splits it into configuration, credential, or memory fragments for separate secure handling.
Validate manifests by applying base-manifest checks and allowlists to catch public domains, hardcoded IPs, and environment-specific values. This process recommends overlayed configurations to prevent misconfigured deployments and maintain base integrity.
Yes, mixed text inputs are split by classifying content into configuration values, credentials, and memory-like facts. Pattern detection identifies each fragment so configuration stays overlay-friendly and observations route to shared agent memory.
Yes, pyyaml is required to parse and process Kubernetes configuration files. The skill uses this dependency to validate manifests against base-manifest rules and perform regex-based content classification on pasted YAML.
Scan git diffs by applying regex pattern detection to flag private keys, API tokens, and hardcoded IPs alongside base-manifest allowlist checks. Discovered credentials route to a password store and all findings appear in a validation report.
Automated content classification may not suit inputs lacking recognizable regex patterns or Kubernetes manifest structures. If content does not match credential patterns, base-manifest rules, or allowlists, the skill cannot accurately route fragments to destinations.
KanakMalpaniA tagline describing your app.
gabergerStore and retrieve agent memory across sessions using AgentDB and ReasoningBank.
rohunjDetect hardcoded secrets in repositories with Gitleaks scans.
SandboxServersHarden Kubernetes and AKS cluster configurations using a security checklist.
