Validate Skill
Run automated validation checks with feedback loops.
Description
The validate skill runs a three-tier validation system:
- •
Pre-flight (Tier 1): Checks before execution
- •Story file exists
- •Budget is available
- •Dependencies are valid
- •
Inter-phase (Tier 2): Checks between agents
- •Code compiles/parses
- •Linting passes
- •Phase transitions are valid
- •
Post-completion (Tier 3): Checks after pipeline
- •Tests pass
- •Types are valid
- •Version is synced
- •Changelog is updated
Usage
code
/validate [story-key] [--tier 1|2|3|all] [--auto-fix] [--json]
Arguments
| Argument | Description | Default |
|---|---|---|
| story-key | Story identifier to validate | Current branch |
| --tier | Validation tier (1, 2, 3, or all) | all |
| --auto-fix | Attempt automatic fixes | false |
| --json | Output as JSON | false |
| --quiet | Minimal output | false |
Examples
Run all validations
code
/validate 3-5 --tier all
Run post-completion checks with auto-fix
code
/validate --tier 3 --auto-fix
Get JSON output for CI
code
/validate 3-5 --json
Prompt
You are running validation checks for the Devflow project.
$ARGUMENTS
Execute the validation loop:
python
import sys
sys.path.insert(0, "tooling/scripts/lib")
from validation_loop import (
create_validation_loop,
run_preflight_validation,
run_post_completion_validation,
ALL_GATES,
LoopContext,
)
# Parse arguments
story_key = "$ARGUMENTS".split()[0] if "$ARGUMENTS" else "validation-check"
tier = "all" # Default to all tiers
# Create context
context = LoopContext(story_key=story_key)
# Run appropriate validation
if tier == "1" or tier == "all":
print("[VALIDATION] Running pre-flight checks...")
report = run_preflight_validation(story_key)
print(report.to_summary())
if tier == "3" or tier == "all":
print("[VALIDATION] Running post-completion checks...")
report = run_post_completion_validation(story_key)
print(report.to_summary())
Report results clearly with:
- •[PASS] for passed gates
- •[WARN] for warnings
- •[FAIL] for failures
- •Suggested fixes when available