vultr/openclaw
1
0
Fork 0
mirror of https://github.com/openclaw/openclaw.git synced 2026-04-22 06:32:00 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix(openai): soften gpt-5 execution bias prompt

Browse Source
This commit is contained in:
Peter Steinberger
2026-04-06 13:36:45 +01:00
parent ef923805f5
commit a2b065b090
2 changed files with 6 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
extensions/openai/index.test.ts
View File

@@ -323,6 +323,9 @@ describe("openai plugin", () => {
expect(OPENAI_GPT5_EXECUTION_BIAS).toContain(
"Do prerequisite lookup or discovery before dependent actions.",
);
expect(OPENAI_GPT5_EXECUTION_BIAS).toContain(
"Stop once the request is fully handled; do not add extra tool calls just for optional polish.",
);
expect(OPENAI_GPT5_OUTPUT_CONTRACT).toContain(
"Return the requested sections only, in the requested order.",
);

5
extensions/openai/prompt-overlay.ts
View File

@@ -45,9 +45,10 @@ export const OPENAI_GPT5_EXECUTION_BIAS = `## Execution Bias
Start the real work in the same turn when the next step is clear.
Do prerequisite lookup or discovery before dependent actions.
If another tool call would likely improve correctness or completeness, keep going instead of stopping at partial progress.
If another tool call is clearly required to handle a requested item or fix correctness, keep going.
Stop once the request is fully handled; do not add extra tool calls just for optional polish.
Multi-part requests stay incomplete until every requested item is handled or clearly marked blocked.
Before the final answer, quickly verify correctness, coverage, formatting, and obvious side effects.`;
Before the final answer, quickly verify correctness, coverage, formatting, and obvious side effects without reopening settled work.`;
export type OpenAIPromptOverlayMode = "friendly" | "off";