Markdown Converter
Agent skill for markdown-converter
Copy-paste prompts for executing phases from your EXECUTION_PLAN.md.
Sign in to like and favorite skills
Copy-paste prompts for executing phases from your EXECUTION_PLAN.md.
Use these prompts when working with documents generated from the greenfield workflow (PRODUCT_SPEC.md, TECHNICAL_SPEC.md, EXECUTION_PLAN.md, AGENTS.md).
Use this when beginning work on a new project or onboarding to an existing execution plan.
Read the following files to understand the structure and purpose of this project: - AGENTS.md (workflow guidelines) - PRODUCT_SPEC.md (what we're building and why) - TECHNICAL_SPEC.md (how it's built technically) - EXECUTION_PLAN.md (tasks and acceptance criteria) Summarize your understanding and confirm you're ready to begin execution.
Use this before starting a new phase to verify prerequisites are met.
I want to execute Phase {N} from EXECUTION_PLAN.md. Before starting: 1. Check the Pre-Phase Setup section — are there unmet prerequisites I need to complete? 2. Verify dependencies — are all prior phases complete? 3. Run a criteria audit for Phase 1 — are acceptance criteria tagged with verification metadata? 4. Ensure verification config is set — if missing, run /configure-verification 5. Review permissions — are there any permissions needed in configuration files for autonomous execution? Report what's ready and what's blocking.
Use this to execute all tasks in a phase autonomously.
I have completed all prerequisites for Phase {N}. Execute all steps and tasks in Phase {N} from EXECUTION_PLAN.md, one at a time. Guidelines: - Read AGENTS.md thoroughly for workflow conventions (git branches, TDD, etc.) - Use the code-verification skill for every task - Do not check back until the phase is complete, unless blocked - Report blockers using the format in AGENTS.md Begin with Step {N}.1.
Use this after completing a phase to verify before proceeding.
Phase {N} is complete. Run the checkpoint criteria from EXECUTION_PLAN.md: 1. Automated checks — run tests, type checking, linting 2. Manual verification — list items for me to verify 3. Summarize what was built in this phase Report results and confirm ready to proceed to Phase {N+1}.
Use these prompts when working with documents generated from the feature workflow (FEATURE_SPEC.md, FEATURE_TECHNICAL_SPEC.md, EXECUTION_PLAN.md, AGENTS_ADDITIONS.md).
Use this when beginning work on a new feature.
Read the following files to understand this feature and how it integrates: - AGENTS.md (workflow guidelines) - FEATURE_SPEC.md (what the feature does and why) - FEATURE_TECHNICAL_SPEC.md (how it integrates technically) - EXECUTION_PLAN.md (tasks and acceptance criteria) Summarize your understanding of the feature and its integration points. Confirm you're ready to begin execution.
Use this before starting a new phase to verify prerequisites are met.
I want to execute Phase {N} from EXECUTION_PLAN.md for this feature. Before starting: 1. Check the Pre-Phase Setup section — are there unmet prerequisites I need to complete? 2. Verify dependencies — are all prior phases complete? 3. Run a criteria audit for Phase 1 — are acceptance criteria tagged with verification metadata? 4. Ensure verification config is set — if missing, run /configure-verification 5. Review integration points — are the existing components we're modifying in the expected state? 6. Check for any permissions needed in configuration files for autonomous execution Report what's ready and what's blocking.
Use this to execute all tasks in a phase autonomously.
I have completed all prerequisites for Phase {N}. Execute all steps and tasks in Phase {N} from EXECUTION_PLAN.md, one at a time. Guidelines: - Read AGENTS.md thoroughly for workflow conventions (git branches, TDD, etc.) - Use the code-verification skill for every task - Pay attention to "Existing Code to Reference" in each task for patterns to follow - Run existing tests after modifications to catch regressions - Do not check back until the phase is complete, unless blocked - Report blockers using the format in AGENTS.md Begin with Step {N}.1.
Use this after completing a phase to verify before proceeding.
Phase {N} is complete. Run the checkpoint criteria from EXECUTION_PLAN.md: 1. Automated checks — run tests (including existing tests), type checking, linting 2. Regression verification — confirm existing functionality still works 3. Manual verification — list items for me to verify 4. Browser verification (if applicable) — verify UI acceptance criteria Report results and confirm ready to proceed to Phase {N+1}.
Use these when execution is interrupted or blocked. Optional: some are also available as slash commands if you've installed the recovery commands.
Use
/phase-analyze to understand what went wrong in a phase before deciding how to proceed.
/phase-analyze 2 # Analyze Phase 2 /phase-analyze # Analyze current/most recent phase
This generates a comprehensive report including:
Use
/phase-rollback when you need to undo work and start fresh.
/phase-rollback 1 # Rollback to end of Phase 1 /phase-rollback 2.1.A # Rollback to just after Task 2.1.A /phase-rollback --last # Undo only the most recent task commit
The command will:
Use
/task-retry when a specific task failed and needs another attempt.
/task-retry 2.1.A # Retry with fresh context /task-retry 2.1.A --fresh # Explicitly start clean /task-retry 2.1.A --alternative # Try a different approach
The
--alternative flag is useful when the standard approach keeps failing. It will:
Use this when a blocker has been resolved and you want to continue.
The blocker for Task {X.Y.Z} has been resolved. Resolution: {describe what was fixed or provided} Continue execution from Task {X.Y.Z}. Pick up where you left off.
Use this when new work is discovered that's outside the current task scope.
While working on Task {X.Y.Z}, I discovered: {description of issue or opportunity} Add this to TODOS.md with appropriate context and priority, then continue with the current task scope. Do not expand scope without approval.
Use this to jump to a specific task (e.g., after manual fixes).
Skip to Task {X.Y.Z} in EXECUTION_PLAN.md. Context: {why we're skipping, what was done manually} Begin execution from this task. Verify any dependencies are met before starting.