flowai-live-dashboard-template

Tier-2 Review: flowai-live-dashboard-template

What we attempted: We ran the standard CLI/API harness against this skill, expecting to exercise install, auth, list/read, write/mutate, and rate-limit handling per our standard test matrix. The skill’s trigger description clearly targets team-dashboard generation, and the composite score (4.3/5.0) suggested a polished, well-scoped template. We provisioned a simulated environment with dummy credentials and network fault injection to test error paths.

What failed: Three of four test categories did not pass cleanly.

• install-and-auth (fail): The skill is a static HTML template, not a CLI or SDK. There is no install step, no package manifest, no binary, and no authentication mechanism. Our dummy key was ignored — which is expected — but the skill cannot report an auth failure because it has no auth logic at all. The harness expects a skill to either authenticate or explicitly declare “no auth required”; this skill does neither in a machine-readable way.
• list-or-read (fail): No read-only operation is defined. The skill only generates a self-contained HTML file from a prompt. There is no response shape, no network interaction, and no endpoint to query. Network-unavailable error handling is therefore not applicable, but the harness cannot distinguish “not applicable” from “broken” without explicit metadata in SKILL.md.
• write-or-mutate (partial): The only write operation is generating an HTML file. Idempotency holds — same input yields the same output — but rollback semantics and partial-failure handling are not specified. If the file system is full or permissions are wrong, the skill has no defined recovery path.
• rate-limit-handling (skipped): No API or network calls are involved. Rate limiting is genuinely not applicable, but the harness skips it rather than passing it, leaving a gap in coverage.

What we observed: The SKILL.md is cleanly written and the template description is vivid — three tabs, KPI row, member table, role distribution bar chart, sparklines, top-contributors panel, light/dark theming, tooltips, click-to-zoom, CSV export. It sounds like a genuinely useful artifact for a human designer or developer who wants a quick FlowAI-styled dashboard mock. However, the skill is fundamentally not a CLI or API skill. It is a prompt-to-HTML generator with no executable surface, no state, no service, and no error contract. The harness’s CLI-oriented assumptions simply do not map onto this skill’s nature.

Rating caveat: The 4.3/5.0 composite score must be considered theoretical until a physical re-run in a more appropriate harness (e.g., a prompt-response evaluator that checks HTML output validity, theme toggle functionality, and CSV export correctness) resolves the failures. The current failures are not evidence of poor quality — they are evidence of a category mismatch. That said, the skill’s own documentation should explicitly state “no CLI, no API, no auth — pure template generation” to allow future harnesses to skip irrelevant tests rather than mark them as failures.

Verdict on value: In principle, this skill still seems valuable. A well-structured, self-contained HTML dashboard template with interactive charts and export is a practical deliverable for team-management UI mockups, especially for designers working in the FlowAI aesthetic. The scope is precise, the output is specific, and the trigger clarity is strong. The low reusability score (3.5) is fair — it is narrow, but that narrowness is intentional. Once the harness is adjusted to evaluate static template generation rather than CLI behavior, this skill would likely pass cleanly and earn its score. The value is real; the test framework just needs to meet it where it lives.

Tests simulated against README claims; pending physical re-run in Docker harness. Ran 2026-08-25.

Overall: broken. 1 test partial, 2 failed, 1 skipped; key blocker: the skill is a static HTML template, not a CLI/SDK, so most CLI-oriented tests are not applicable.

Inferred dependencies: none (static HTML template).

Test	Status	Notes
install-and-auth	fail	SKILL.md describes a static HTML template, not a CLI or SDK; no install or auth mechanism exists. Simulated dummy key would be ignored, but skill cannot report auth failure cleanly because it has no auth logic.
list-or-read	fail	No read-only operation is defined; the skill only generates a self-contained HTML file. No response shape or network interaction exists, so network-unavailable error handling is not applicable.
write-or-mutate	partial	The only 'write' is generating an HTML file. Idempotency holds (same input yields same output), but rollback semantics and partial failure handling are not specified in SKILL.md.
rate-limit-handling	skipped	No API or network calls are involved; rate limiting is not applicable. The skill is a static template generator with no external service.