Skip to content

fix: load_skill handles dir paths; validate evolved_full not body #51

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
shilei103-afk wants to merge 1 commit into
base: main
Choose a base branch
from
Open

fix: load_skill handles dir paths; validate evolved_full not body #51

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
37 changes: 21 additions & 16 deletions evolution/skills/evolve_skill.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -186,7 +186,7 @@ def evolve(

# ── 7. Validate evolved skill ───────────────────────────────────────
console.print(f"\n[bold]Validating evolved skill[/bold]")
evolved_constraints = validator.validate_all(evolved_body, "skill", baseline_text=skill["body"])
evolved_constraints = validator.validate_all(evolved_full, "skill", baseline_text=skill["raw"])
all_pass = True
for c in evolved_constraints:
icon = "✓" if c.passed else "✗"
Expand All @@ -204,27 +204,32 @@ def evolve(
console.print(f" Saved failed variant to {output_path}")
return

# ── 8. Evaluate on holdout set ──────────────────────────────────────
# ── 8. Evaluate on holdout set ───────────────────────────────────────
console.print(f"\n[bold]Evaluating on holdout set ({len(dataset.holdout)} examples)[/bold]")

holdout_examples = dataset.to_dspy_examples("holdout")

baseline_scores = []
evolved_scores = []
for ex in holdout_examples:
# Score baseline
with dspy.context(lm=lm):
baseline_pred = baseline_module(task_input=ex.task_input)
baseline_score = skill_fitness_metric(ex, baseline_pred)
baseline_scores.append(baseline_score)

evolved_pred = optimized_module(task_input=ex.task_input)
evolved_score = skill_fitness_metric(ex, evolved_pred)
evolved_scores.append(evolved_score)

avg_baseline = sum(baseline_scores) / max(1, len(baseline_scores))
avg_evolved = sum(evolved_scores) / max(1, len(evolved_scores))
improvement = avg_evolved - avg_baseline
try:
for ex in holdout_examples:
# Score baseline
with dspy.context(lm=lm):
baseline_pred = baseline_module(task_input=ex.task_input)
baseline_score = skill_fitness_metric(ex, baseline_pred)
baseline_scores.append(baseline_score)

evolved_pred = optimized_module(task_input=ex.task_input)
evolved_score = skill_fitness_metric(ex, evolved_pred)
evolved_scores.append(evolved_score)
except Exception as e:
console.print(f"[yellow]⚠ Holdout evaluation failed (DSPy adapter error): {e}[/yellow]")
console.print(" Constraints passed — skill is valid but holdout scores unavailable.")
avg_baseline = avg_evolved = improvement = 0.0
else:
avg_baseline = sum(baseline_scores) / max(1, len(baseline_scores))
avg_evolved = sum(evolved_scores) / max(1, len(evolved_scores))
improvement = avg_evolved - avg_baseline

# ── 9. Report results ───────────────────────────────────────────────
table = Table(title="Evolution Results")
Expand Down
7 changes: 7 additions & 0 deletions evolution/skills/skill_module.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -25,6 +25,13 @@ def load_skill(skill_path: Path) -> dict:
"description": str,
}
"""
# If skill_path is a directory, find SKILL.md inside it
if skill_path.is_dir():
skill_md = skill_path / "SKILL.md"
if not skill_md.exists():
raise FileNotFoundError(f"No SKILL.md found in {skill_path}")
skill_path = skill_md

raw = skill_path.read_text()

# Parse YAML frontmatter
Expand Down