Skip to content

ericrisco/rsc-harness / skills/landing-copy

landing-copy

Verified

Use when writing the words on a single conversion page — the hero headline and subhead, the offer, the proof and testimonials, and one primary CTA — or diagnosing a page that gets traffic and does not convert, as a copy problem rather than a layout one. NOT the paid ad that drives the click (that is `ads`), NOT the visual layout the words sit in (that is `design`), NOT the experiment that picks the winning variant (that is `ab-testing`).

FileKindSize
SKILL.mdskill md12.2 KB
evals/README.mdmarkdown1.6 KB
evals/cases.yamlconfig5.5 KB
references/frameworks.mdmarkdown3.4 KB
references/page-skeleton.mdmarkdown2.5 KB
references/vsl-page-handoff.mdmarkdown1.4 KB
scripts/verify.shexecutablescript6.3 KB

evals/README.md

View on GitHub
# Eval harness — `landing-copy` skill

These cases are read and run by the catalog's agent harness (a Claude Code agent
with the full skill catalog loaded for routing) plus a human spot-check — there
is no bundled shell runner, because both triggering and copy quality are
judgement calls. `scripts/verify.sh` is a separate, narrower static linter for a
*generated copy file*, not a runner for these cases.

`should_trigger` asserts that the description + body would make a router load
`landing-copy` for that prompt (run each cold, 3–5 trials, score the fraction
that load the skill). `should_not_trigger` asserts a near-miss routes to the
named real sibling instead (`ads`, `design`, `nextjs`, `ab-testing`,
`brand-voice`) — these are the known traps: a "landing page" prompt that is
actually the ad, the pixels, the build, or the experiment. `capability` is a
rubric-scored generation: run the scenario with and without the skill loaded and
grade each output against `must_include` (one point per item); the skill passes
if WITH covers >= 80% and clearly beats WITHOUT. Report raw fractions, not a
rounded pass, and re-run after any edit to `SKILL.md` or the references.

## VSL integration validation — 2026-10-31

The added offer-revision scenario and source-grounding repairs were statically reviewed.
The proof-marker regression tests exercise demonstration-only evidence and paths with spaces.
The revised skill was applied to the product README. These checks are not a fresh A/B measurement
of landing-copy quality; the capability comparisons described above remain a separate evaluation.
finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the workStars: 322kUpdated Claude CodeCodexVerifiedInstall
systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixesStars: 322kUpdated Claude CodeCodexVerifiedInstall
test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation codeStars: 322kUpdated Claude CodeCodexVerifiedInstall
skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.Stars: 198kUpdated Claude CodeCodexVerifiedInstall