Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →withkynam/vc-plan-discovery
Discover related plans for the current task: same feature folder full depth, other features active-only, general-plans active. Like vc-context-discovery but for plan artifacts.
Passed 12 of 17 test cases with the skill, and 3 of 17 without it.
and 10 more in the full report
13 fixed3 passed either way5 still failing1 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 13 test cases the skill fixed. Same model, same prompt, with and without the skill.
We want to discover planning documents specifically for the checkout component using the discovery script. Which option specifies the domain targeted?
| Checked | Without skill | With skill |
|---|---|---|
| Uses the --feature flag with checkout as the parameter value | Missed | Passed |
The instructions your model reads when the skill loads.
Output style: Follow
process/development-protocols/communication-standards.md— answer-first, plain language, no unexplained jargon, TL;DR on long responses.
Primary method — run the auto-discovery script. It applies the Scope Rules below
(same-feature full depth; other features active-only; general-plans active always) and
extracts ONLY the leading YAML frontmatter block of each .md file (no whole-file reads):
node .claude/skills/vc-plan-discovery/scripts/discover-plans.mjs [--feature <name>] [--json]
It groups output into Active Plans / Backlog / Completed / Reports / References (each line:
- [name]: description (path)) and prints a trailing Found N active, N backlog, N completed, N reports, N references line. It never throws on a missing root and exits 0
unless given a bad flag. Use --json for a machine-readable object. Prefer this over the
manual scan below — it is deterministic and avoids loading huge files into context.
Per task-folder artefact colocation, the script scans one level into each
{slug}_{date}/ task folder and surfaces co-located _PLAN_/_REPORT_/_REF_ artefacts;
the sibling reports//references/ dirs are deprecated and only hold legacy artefacts.
The manual scan in Scope Rules below is the FALLBACK for when the script fails.
Surface all plans relevant to the current task so agents have full plan context — what was tried, what is in progress, what is deferred, and what reports and references exist — before doing any phase work.
active/, backlog/, completed/, plus any legacy sibling reports/, references/ — surface every file with frontmatteractive/ only — surface plans whose frontmatter description or feature field matches the task domaingeneral-plans/active/: always scangeneral-plans/completed/, legacy reports/, references/: scan only when same-feature folder is not identifiedPer task-folder artefact colocation, expect every current artefact (plan, spec, reports, references) INSIDE its {slug}_{date}/ task folder — scan one level into each task folder. The sibling reports//references/ dirs are deprecated and only hold legacy artefacts.
test cases passed, vs 3 / 17 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.