oprc-procedure-generation

name	oprc - Procedure Generation
description	Generates step-by-step procedures with dependency ordering, decision points, and failure handling. Takes a goal and produces a complete, testable procedure.
output	{"format":"prose"}

Procedure Generation

Input: $ARGUMENTS

---

Step 1: Identify the Goal

State what the procedure accomplishes when followed correctly.

GOAL: [what successful completion looks like]
SCOPE: [what the procedure covers / what it does not cover]
PREREQUISITES: [what must be true before starting]
ACTOR: [who performs this procedure]

---

Step 2: List All Steps in Dependency Order

Enumerate every step required to reach the goal. Order by dependency, not by importance.

Rules:

• Each step should be a single action, not a compound task
• If a step contains "and", consider splitting it
• Steps must be in an order where each step's inputs are available from prior steps
• Number steps sequentially

STEPS:
1. [action verb] [what] [with what inputs]
2. [action verb] [what] [with what inputs]
3. [action verb] [what] [with what inputs]
...

---

Step 3: Specify Inputs and Outputs

For each step, define what goes in and what comes out. This makes the procedure auditable and debuggable.

STEP [N]: [step description]
  INPUT: [what this step needs]
  OUTPUT: [what this step produces]
  DONE WHEN: [how to verify this step succeeded]

Rules:

• Every output should be consumed by a later step or be a final deliverable
• Every input should come from a prior step's output or from the prerequisites
• If an input has no source, the procedure has a gap

---

Step 4: Add Decision Points

Identify where the procedure branches based on conditions.

STEP [N]: Check [condition]
  IF [condition A]: proceed to Step [X]
  IF [condition B]: proceed to Step [Y]
  IF [neither]: [what to do — escalate, retry, abort]

Rules:

• Decision points must be binary or have a small number of clear branches
• Every branch must eventually rejoin the main flow or reach a defined endpoint
• Include a default/fallback for unexpected conditions

---

Step 5: Mental Walkthrough

Test the procedure by walking through it with a concrete example.

TEST SCENARIO: [a specific realistic case]

WALKTHROUGH:
- Step 1: [what happens with this input]
- Step 2: [what happens next]
- ...
- Result: [what the procedure produces]

ISSUES FOUND:
- [any gaps, ambiguities, or missing steps discovered]

Fix any issues found before proceeding.

---

Step 6: Add Failure Handling

For each step that can fail, define what to do.

Step	Failure mode	Response
[N]	[what could go wrong]	[retry / skip / escalate / abort]
[N]	[what could go wrong]	[retry / skip / escalate / abort]

Rules:

• Every step that interacts with external systems needs failure handling
• Define maximum retry counts where applicable
• Identify the point of no return (if any) after which rollback is not possible
• Include a full abort procedure if the process must be abandoned mid-way

---

Step 7: Deliver

Present the complete procedure with:

• Goal and prerequisites at the top
• Numbered steps with inputs, outputs, and verification criteria
• Decision points clearly marked
• Failure handling included inline or as an appendix
• An estimated time or effort for the full procedure (if applicable)

PROCEDURE: [name]
GOAL: [what it accomplishes]
PREREQUISITES: [what must be true before starting]
ESTIMATED EFFORT: [time or effort estimate]

[The complete procedure]

---

Integration

Use with:

• /de -> Design the system that the procedure operates within
• /to -> Break the procedure into a task outline for execution
• /oart -> Package the procedure as a reusable template
• /pv -> Validate the procedure against edge cases