The Effort Scale
Quantifying cognitive load and architectural complexity on a 1-to-10 scale.
In Fireskills, Effort is a numerical rating from 1 to 10 assigned to every specification. Unlike traditional agile points or hour estimates, Effort explicitly measures cognitive reasoning load, architectural risk, and execution complexity — never calendar days or deadlines.
The 4 Complexity Profiles
| Effort Range | Profile | Typical Scope | Agent Reasoning Profile |
|---|---|---|---|
| 1 – 2 | light | Minor adjustments, copy edits, localized bug fixes, single-component tweaks. | Fast execution, minimal architectural reasoning needed. |
| 3 – 6 | standard | Typical feature slices: new tRPC procedure, Drizzle table addition, CRUD page, or UI form. | Standard multi-step reasoning, requires domain specialist. |
| 7 – 8 | high | Cross-layer features: Worker bindings, queue consumer, billing integration, complex RBAC. | Deep architectural planning, rigorous multi-file verification. |
| 9 – 10 | maximum | Core infrastructure refactor, authentication system overhaul, data migration at scale. | Maximum reasoning budget, adversarial evaluation (fireskill-the-fool). |
Metadata Header Fields
The specification header tracks three tightly linked effort fields:
| Campo | Valor |
| --- | --- |
| Effort | 4 |
| Effort updated at | 2026-09-18T16:15:00.000Z |
| Effort rationale | Adds a Stripe webhook handler and idempotent subscription updates |
Updating Effort via CLI
Effort should be updated whenever new discovery, task planning, or implementation reveals material complexity that was previously unknown.
Use the fireskill CLI command:
bun run fireskill effort <id> <1-10> --reason "<rationale>"
Example:
bun run fireskill effort 0042 6 --reason "Requires custom Drizzle migration and Cloudflare KV session cache invalidation"
The command updates Effort, sets Effort updated at to the current ISO 8601 timestamp and records the Effort rationale in the header. The identifier is the spec number or the full <NNNN>-<slug>; a score outside 1–10, or an empty reason, is rejected.
How to use it
The score is a guide for how much care a slice needs. It does not start any skill by itself. As a rule of thumb, for a spec with Effort 7 or more, ask for an adversarial review ($fireskill-the-fool) before approving the Definition Gate, and for a collegiate decision ($fireskill-the-jury) on irreversible architecture choices. $fireskill-interviewer re-evaluates the score when a conversation changes the capability required.
Ready to build and launch your SaaS?
Get 100% full source code ownership, zero proprietary wrappers, and architecture engineered for millions of requests on Cloudflare.