Guardrails Learning Loop
When a check fails twice the same way, append a guardrail sign to .ralph/guardrails.md so the next iteration avoids repeating it.
// parameters
// run_this_loop
Copy the kickoff into your agent first. The blueprint below explains how the loop evaluates progress and when it should stop.
Use loop copies the kickoff prompt. Share copies the loop link. Open in Cursor / Claude only pre-fill that prompt — they do not install hook files. Download loop saves the loop as guardrails-learning-loop.md.
Start the "Guardrails Learning Loop" loop.
Goal: tests and lint pass without repeating prior failure patterns
Max iterations: 12
Between iterations run: npm test && npm run lint
Exit when: all checks pass
Step 1: Read .ralph/guardrails.md, run checks, and if a failure repeats, add a sign before fixing.
Self-pace this loop. After each iteration, run the check command, read the output, and only continue if the exit condition is not met. Stop when the exit condition passes or max iterations is reached. Give a short status update each pass.// blueprint
The loop contract: target outcome, execution sequence, checks, stopping condition, and safety constraints.
tests and lint pass without repeating prior failure patterns
// loop_sequence
Trigger
Manual start
Actions block
Feedback gate
Evaluate: npm test && npm run lint
Run the check, read the output, then decide whether the loop is complete.
Check failed
Repeat actions with the latest feedback.
Success
Checks pass and no repeated failure pattern remains
// steps
- 1Read guardrails
Read .ralph/guardrails.md before any work. Treat each sign as a hard constraint.
- 2Run checks
Run the project's backpressure commands.
npm test && npm run lint - 3Record failure sign
If the same error occurred before, append a concise sign to guardrails.md: what failed and how to avoid it.
- 4Fix with guardrails
Apply a fix that respects all guardrail signs. Do not repeat prior failed approaches.