Skip to content

Agentic Workflow

Use this workflow when an AI agent should retrieve issue/PR context before proposing or implementing changes.

Setup

Initialize the canonical skill scaffold:

99problems skill init

Optional custom location:

99problems skill init --path ~/.agents/skills

The generated 99problems skill follows the Agent Skills standard.

How Agents Use It

  • Explicit skill invocation: invoke it directly with $99problems in your prompt.
llm-prompt> $99problems summarize related Jira issues and GitHub PRs for topic "document generation redesign"
  • Implicit skill invocation: assign a task that requires issue/PR context retrieval and let the agent activate the skill automatically.
llm-prompt> Create an overview of open bugs and cross-reference them with active PRs in owner/repo.

Prototypical Tasks

  • Build an overview of work on topic X across issues and PRs.
  • Cross-reference issues with linked PRs and summarize status.
  • Create a bug overview for a repository/project (open vs closed, high-priority focus).
  • Estimate progress from issue/PR states and recent updates.
  • Surface blockers/blocked-by links and dependency hotspots.
  • Summarize key decisions from discussion threads for handoff.

Next References