Content
82%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The body is a strong operational document: fully executable commands for all three phases, explicit prerequisites with documented fallbacks, idempotency and replay semantics, and a pre-entry review checklist. Weaknesses are modest — repeated safety disclaimers and time-sensitive regulatory dates hurt conciseness slightly, no explicit error-recovery loops in the workflow, and the large smoke-test runbook in the bundle is not surfaced in the Resources section.
Suggestions
Consolidate the no-order-routing safety statement into one authoritative location (e.g., the Overview or a dedicated Safety section) instead of repeating it in the Overview, "Do NOT invoke", and Phase 3 sections, and move the FINRA effective-date detail into references/short_risk_management.md.
Add references/smoke_test_runbook.md and the smoke_universe_*.csv fixtures to the Resources section so the bundle's testing artifacts are discoverable from the main file.
Add explicit error-recovery guidance for each phase (e.g., what to check when the FMP call fails, when output fails schema validation, or when Alpaca data is unavailable mid-session) to close the feedback-loop gap in workflow clarity.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense and operational — full commands with flags, no explanations of concepts Claude already knows, and tight section structure. It is not a 5 because of minor trimmable material: the no-order-routing safety message is repeated three times (Overview, "Do NOT invoke", and the Phase 3 disclaimer), the FINRA date detail ("effective 2026-06-04, with broker phase-in allowed through 2027-10-20") is time-sensitive content outside a deprecated section, and legacy-compat notes ("earnings_within_2d is kept for backward compatibility") pad the earnings section. It is clearly above the 3 anchor since the bulk is efficient and every section carries operational weight. | 4 / 5 |
Actionability | All three phases have copy-paste-ready commands with complete flags, input/output paths, and account parameters (e.g., "python3 skills/parabolic-short-trade-planner/scripts/screen_parabolic.py --mode safe_largecap --as-of 2026-04-30 --output-dir reports/"), plus env-var setup, fallback behavior without API keys, and fixture-based testing paths ("--dry-run --fixture <path>", "--bars-source fixture") that cover the common cases including API-less runs. Output file names and key fields are enumerated so the operator knows exactly what to inspect. This matches the 5 anchor; the only trivial nit (no literal `pip install -r requirements.txt` command) does not constitute a gap. | 5 / 5 |
Workflow Clarity | The three-phase workflow is clearly sequenced with numbered steps, per-phase prerequisites (confirm FMP_API_KEY / ALPACA keys), documented fallbacks when prerequisites are missing, and an explicit pre-entry review checklist (plan_status / blocking_manual_reasons / advisory_manual_reasons). It falls short of the 5 anchor because there are no explicit validate→fix→retry feedback loops (e.g., what to do when a run fails or output fails schema validation is only implied via the referenced test suite), while it exceeds the 3 anchor because checkpoints are present and explicit at each phase, and the detection-only design means destructive-operation risk is absent by construction. | 4 / 5 |
Progressive Disclosure | Good structure overall: the body keeps operational workflow inline and delegates domain depth to five clearly-signaled, one-level-deep reference files with one-line descriptions, and every referenced bundle path (references/*.md, scripts/tests/fixtures/dry_run_minimal.json, scripts/tests/fixtures/intraday_bars/, tests/test_schema_contract.py, tests/test_monitor_intraday_smoke.py) exists. It is not a 5 because the actual bundle contains discoverable content not navigable from SKILL.md — notably the 21KB references/smoke_test_runbook.md and the two smoke_universe_*.csv files are absent from the Resources section, leaving a substantial artifact undiscoverable. | 4 / 5 |
Total | 17 / 20 Passed |