askill
promptcontext

promptcontextSafety 90Repository

Context recommendation system for reducing action uncertainty. Use PROACTIVELY when user mentions keywords like 'use this kubeconfig', 'our goal is', 'debug this', 'in production', 'deploy to', or when uncertainty about environment, cluster, namespace, or AWS profile drops below 80%.

0 stars
1.2k downloads
Updated 2/18/2026

Package Files

Loading files...
SKILL.md

Prompt Context Recommendation

This skill provides guidance for when and how to recommend setting operational context to reduce uncertainty in actions.

Uncertainty Threshold

Critical Rule: When confidence in any of these is below 80%, PROACTIVELY recommend setting context:

Context TypeSigns of UncertaintyRecommendation
KubernetesUnknown cluster, namespace, or kubeconfig/prompt:context set "k8s <cluster> namespace <ns>"
AWSUnknown profile or region/prompt:context set "aws profile <name> region <region>"
GitWrong branch assumptions, repo confusion/prompt:context set "git <branch>"
ProjectUnclear which project or service/prompt:context set "note: working on <project>"
EnvironmentUnknown dev/staging/prod/prompt:context set "env ENV=<value>"

When to Trigger

Keyword-Based Triggers (PROACTIVE)

When user message contains these keywords/patterns, immediately recommend setting context:

Keyword PatternRecommended ActionExample Command
use this kubeconfig / kubeconfig file: / ~/.kube/Set k8s config/prompt:context set "k8s <cluster> --kubeconfig <path>"
our goal is / we're working on / the objectiveSet project note/prompt:context set "note: <captured goal>"
debug this / troubleshoot / investigateSet debug context/prompt:context set "note: debugging <component>"
in production / prod environmentSet prod context/prompt:context set "env production; note: PROD - be careful"
staging / in stagingSet staging context/prompt:context set "env staging"
deploy to / deployingSet deployment target/prompt:context set "note: deploying to <target>"
AWS account / aws profileSet AWS profile/prompt:context set "aws profile <name>"
this cluster / cluster namedSet k8s cluster/prompt:context set "k8s <cluster-name>"
namespace / in nsSet k8s namespace/prompt:context set "k8s <cluster> namespace <ns>"
this repo / this project / codebaseSet project note/prompt:context set "note: working on <project-name>"

Keyword Trigger Examples:

  1. User says: "use this kubeconfig file: ~/.kube/nexus-dev"

    Immediately recommend:

    /prompt:context set "k8s nexus-dev --kubeconfig ~/.kube/nexus-dev"
    
  2. User says: "our goal is to fix the authentication bug in the API gateway"

    Immediately recommend:

    /prompt:context set "note: fixing auth bug in API gateway"
    
  3. User says: "debug this failing deployment"

    Immediately recommend:

    /prompt:context set "note: debugging failing deployment"
    

    Then ask: Which cluster and namespace?

  4. User says: "we're working on the payment-service microservice"

    Immediately recommend:

    /prompt:context set "note: working on payment-service"
    

Automatic Triggers (Uncertainty < 80%)

  1. Kubernetes operations without known context:

    • "Apply this manifest" - Which cluster? Which namespace?
    • "Check pod status" - Which cluster?
    • "Scale deployment" - Where?
  2. AWS operations without profile/region:

    • "Deploy to S3" - Which account? Which region?
    • "Create Lambda" - What profile?
    • "Update IAM" - Production or development?
  3. Multi-environment codebases:

    • Operations that could affect dev, staging, or production
    • Database migrations, deployments, configuration changes
  4. Ambiguous commands:

    • User says "deploy" without specifying where
    • "Run tests" in a multi-service repo
    • "Check logs" without service/namespace context

Response Patterns

Pattern 1: Keyword Detection (Immediate)

When keywords are detected, respond immediately with context suggestion:

I noticed you mentioned [keyword]. Let me capture that context:

/prompt:context set "[captured context]"


[Continue with the task using this context]

Examples:

  • Keyword detected: "use this kubeconfig file: ~/.kube/nexus-dev"

    I'll set up the kubeconfig context:

    /prompt:context set "k8s nexus-dev --kubeconfig ~/.kube/nexus-dev"
    

    Now proceeding with your request using this cluster...

  • Keyword detected: "our goal is to migrate the database"

    Capturing your goal:

    /prompt:context set "note: goal - migrate the database"
    

    I'll keep this objective in mind. Which database and target environment?

Pattern 2: Uncertainty Detection (Threshold)

When uncertainty is detected, respond with:

**Context Recommendation**

I noticed uncertainty about the operational context. To ensure accurate actions, consider setting:

/prompt:context set ""


**Current uncertainty:**
- [List specific unknowns: cluster, namespace, profile, etc.]

**Why this matters:**
- [Brief explanation of potential risks without context]

Would you like to set the context, or should I proceed with assumptions?

Quick Context Commands

# View current context
/prompt:context

# Set Kubernetes context
/prompt:context set "k8s prod-cluster namespace api"

# Set AWS context
/prompt:context set "aws profile production region us-west-2"

# Set project note
/prompt:context set "note: working on payment-service"

# Combined context
/prompt:context set "k8s dev-cluster; aws staging; note: testing feature X"

# Clear context
/prompt:context clear all

Uncertainty Calculation

Estimate confidence based on:

FactorHigh Confidence (>80%)Low Confidence (<80%)
ClusterExplicitly stated or single clusterMultiple possible, none specified
NamespaceStated or default known safeCould be any of several
AWS ProfileStated or obvious from taskMultiple accounts possible
RegionStated or single-region setupMulti-region, none specified
EnvironmentClear dev/staging/prodCould be any

Examples

Example 1: Kubernetes Uncertainty

User: "Scale the api deployment to 5 replicas"

Uncertainty: Cluster and namespace unknown Confidence: ~40%

Response:

Context Recommendation

I'm uncertain about the target cluster and namespace for scaling the api deployment.

/prompt:context set "k8s <cluster-name> namespace <namespace>"

Without context, I might scale the wrong deployment in the wrong environment. Would you like to set the context first?

Example 2: AWS Uncertainty

User: "Upload this file to S3"

Uncertainty: AWS profile and bucket unknown Confidence: ~50%

Response:

Context Recommendation

I need to know which AWS account and bucket to use.

/prompt:context set "aws profile <profile-name> region <region>"

Which S3 bucket should I upload to?

Example 3: High Confidence (No Recommendation Needed)

User: "Scale the api deployment to 5 replicas in the prod-cluster api namespace"

Confidence: 95% - All context explicitly provided Action: Proceed directly without context recommendation

Integration with Other Skills

  • kubernetes-expert: Use promptcontext before k8s operations
  • cloud-architect: Set AWS context before infrastructure changes
  • deployment-engineer: Set full context before deployments

Best Practices

  1. Ask early: Recommend context at start of ambiguous tasks
  2. Don't over-ask: If context is clear, don't prompt
  3. Remember context: Once set, context persists - don't re-ask
  4. Offer defaults: Suggest common contexts as options
  5. Explain risks: Help users understand why context matters

Install

Download ZIP
Requires askill CLI v1.0+

AI Quality Score

85/100Analyzed 2/23/2026

Well-structured skill with excellent actionability and clarity. Provides comprehensive guidance on context recommendation with specific commands, triggers, and response patterns. Penalized for internal-only signals (deeply nested path in plugin skills folder, plugin-specific /prompt:context commands). Despite being internal, the skill demonstrates high quality in documentation structure and could serve as a reference for similar context-management patterns.

90
95
60
95
95

Metadata

Licenseunknown
Version-
Updated2/18/2026
Publisherzach-source

Tags

apici-cddatabasepromptingsecurity