chore(scripts): clarify refactor skill loading instruction in codex loop
This commit is contained in:
@@ -11,7 +11,7 @@ CODEX_MODEL="${CODEX_MODEL:-gpt-5.3-codex}"
|
|||||||
CODEX_REASONING_EFFORT="${CODEX_REASONING_EFFORT:-medium}"
|
CODEX_REASONING_EFFORT="${CODEX_REASONING_EFFORT:-medium}"
|
||||||
|
|
||||||
read -r -d '' PROMPT <<'EOF' || true
|
read -r -d '' PROMPT <<'EOF' || true
|
||||||
1. load the refactor skill
|
1. load the refactor skill and read its SKILL.md fully from top to bottom (do not skim or partially read)
|
||||||
2. check previous commits to see what other agents have worked on
|
2. check previous commits to see what other agents have worked on
|
||||||
3. identify and work on a single improvement based on the refactor skill (can be a file system reorg, file splitting, refactor, add a new test, harden a test, deflakify a test, fix a test, improve control flow, remove unused code)
|
3. identify and work on a single improvement based on the refactor skill (can be a file system reorg, file splitting, refactor, add a new test, harden a test, deflakify a test, fix a test, improve control flow, remove unused code)
|
||||||
4. commit with a description of what was done, your reasoning, and document accomplishments or challenges for the next agent
|
4. commit with a description of what was done, your reasoning, and document accomplishments or challenges for the next agent
|
||||||
|
|||||||
Reference in New Issue
Block a user