docs: rewrite prompt techniques, add nondet rules, rename nav#370
Merged
MuncleUscles merged 2 commits intomainfrom Mar 20, 2026
Merged
docs: rewrite prompt techniques, add nondet rules, rename nav#370MuncleUscles merged 2 commits intomainfrom
MuncleUscles merged 2 commits intomainfrom
Conversation
✅ Deploy Preview for genlayer-docs ready!
To edit notification comments on pull requests, go to your Netlify project configuration. |
Contributor
|
Caution Review failedThe pull request is closed. ℹ️ Recent review info⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (5)
📝 WalkthroughWalkthroughUpdated navigation metadata labels and substantially revised intelligent-contract documentation: renamed a couple of menu entries, replaced the “Crafting Prompts” doc with a new “Prompt & Data Techniques” focused on LLM-grounding and deterministic checks, and added nondeterminism placement rules; validator setup example versions were updated. Changes
Sequence Diagram(s)sequenceDiagram
participant Dev as Developer (writes prompts/rules)
participant LLM as LLM (generates checks)
participant Sandbox as Sandbox Eval (deterministic)
participant API as External API (leader/validator)
participant Contract as Compliance Prompt / Contract
Dev->>LLM: send rule-generation prompt
LLM->>Dev: return JSON-formatted check expressions
Dev->>Sandbox: pass JSON checks for deterministic eval
Sandbox->>API: fetch deterministic fields (independent leader/validator calls)
API-->>Sandbox: return normalized deterministic fields
Sandbox->>Sandbox: evaluate checks -> produce verified results
Sandbox->>Contract: inject verified "ground truth" into compliance prompt
Contract->>LLM: request final compliance judgement (with ground truth)
LLM-->>Contract: final, structured decision
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~20 minutes Possibly related PRs
Suggested reviewers
Poem
🚥 Pre-merge checks | ✅ 3✅ Passed checks (3 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
📝 Coding Plan
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
…ections - Rewrite crafting-prompts.mdx → "Prompt & Data Techniques" with 5 concrete techniques: JSON responses, stable field extraction, derived status comparison, LLM grounding with programmatic eval, error classification - Add "What Goes Inside vs Outside" section to non-determinism.mdx explaining what must/cannot happen inside nondet blocks (storage writes, contract calls, message emission, nesting) - Rename "Decentralized Applications" → "Frontend & SDK Integration" in nav
MuncleUscles
force-pushed
the
docs/improve-prompt-techniques-and-nondet-rules
branch
from
b66fc01 to
9dadc20
Compare
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 2
🧹 Nitpick comments (1)
pages/developers/intelligent-contracts/crafting-prompts.mdx (1)
32-32: Small wording polish for docs quality.Use “Markdown” (proper noun) at Line [32], and “error-handling” (hyphenated compound adjective) at Line [198].
Also applies to: 198-198
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@pages/developers/intelligent-contracts/crafting-prompts.mdx` at line 32, Update the wording in the sentence that currently reads 'Without `response_format="json"`, LLMs may wrap output in markdown code fences, add commentary, or return malformed JSON. With it, you get a parsed dict directly.' to use the proper noun "Markdown" (capitalized). Also change the compound adjective at the other location that references error handling to the hyphenated form "error-handling". Locate these phrases in the document by searching for the exact sentence about response_format="json" and the occurrence of "error handling" (or similar) and apply the two small wording edits.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@pages/developers/intelligent-contracts/features/non-determinism.mdx`:
- Around line 21-31: The example references validator_fn but never defines it;
add a validator function (e.g., validator_fn) alongside leader_fn inside
fetch_price that accepts the leader result (or the raw response) and returns a
boolean indicating validity (for example by verifying parse_price(result)
succeeds and the price is within expected bounds), then pass that validator_fn
to gl.vm.run_nondet_unsafe; ensure the symbols match (leader_fn, validator_fn,
fetch_price, gl.vm.run_nondet_unsafe) so the snippet is runnable.
- Around line 54-58: The validator_fn currently divides by leaders_res.calldata
which can be zero; update validator_fn to guard against divide-by-zero by first
checking if leaders_res.calldata is zero (or near-zero) before performing the
ratio compute — e.g., if leaders_res.calldata == 0 (or abs(leaders_res.calldata)
< small_epsilon) return comparison using direct equality or a tolerance-based
absolute difference, otherwise perform the existing relative check using
abs(leaders_res.calldata - my_price) / leaders_res.calldata <= 0.02; reference
validator_fn, leader_fn, and leaders_res.calldata when implementing the guard.
---
Nitpick comments:
In `@pages/developers/intelligent-contracts/crafting-prompts.mdx`:
- Line 32: Update the wording in the sentence that currently reads 'Without
`response_format="json"`, LLMs may wrap output in markdown code fences, add
commentary, or return malformed JSON. With it, you get a parsed dict directly.'
to use the proper noun "Markdown" (capitalized). Also change the compound
adjective at the other location that references error handling to the hyphenated
form "error-handling". Locate these phrases in the document by searching for the
exact sentence about response_format="json" and the occurrence of "error
handling" (or similar) and apply the two small wording edits.
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: e0495290-6eeb-468b-8ccd-0718e9177841
📒 Files selected for processing (4)
pages/developers/_meta.jsonpages/developers/intelligent-contracts/_meta.jsonpages/developers/intelligent-contracts/crafting-prompts.mdxpages/developers/intelligent-contracts/features/non-determinism.mdx
Comment on lines
+54
to
+58
| def validator_fn(leaders_res) -> bool: | ||
| if not isinstance(leaders_res, gl.vm.Return): | ||
| return False | ||
| my_price = leader_fn() | ||
| return abs(leaders_res.calldata - my_price) / leaders_res.calldata <= 0.02 |
Contributor
There was a problem hiding this comment.
Guard against divide-by-zero in validator ratio check.
Line [58] divides by leaders_res.calldata; if that value is 0, this example throws.
Suggested fix
- return abs(leaders_res.calldata - my_price) / leaders_res.calldata <= 0.02
+ denom = max(abs(leaders_res.calldata), 1e-9)
+ return abs(leaders_res.calldata - my_price) / denom <= 0.02📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| def validator_fn(leaders_res) -> bool: | |
| if not isinstance(leaders_res, gl.vm.Return): | |
| return False | |
| my_price = leader_fn() | |
| return abs(leaders_res.calldata - my_price) / leaders_res.calldata <= 0.02 | |
| def validator_fn(leaders_res) -> bool: | |
| if not isinstance(leaders_res, gl.vm.Return): | |
| return False | |
| my_price = leader_fn() | |
| denom = max(abs(leaders_res.calldata), 1e-9) | |
| return abs(leaders_res.calldata - my_price) / denom <= 0.02 |
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@pages/developers/intelligent-contracts/features/non-determinism.mdx` around
lines 54 - 58, The validator_fn currently divides by leaders_res.calldata which
can be zero; update validator_fn to guard against divide-by-zero by first
checking if leaders_res.calldata is zero (or near-zero) before performing the
ratio compute — e.g., if leaders_res.calldata == 0 (or abs(leaders_res.calldata)
< small_epsilon) return comparison using direct equality or a tolerance-based
absolute difference, otherwise perform the existing relative check using
abs(leaders_res.calldata - my_price) / leaders_res.calldata <= 0.02; reference
validator_fn, leader_fn, and leaders_res.calldata when implementing the guard.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
response_format="json")spawn_sandbox)[EXPECTED]vs[EXTERNAL])Test plan
python scripts/lint-code-examples.pySummary by CodeRabbit