Install any skill in seconds. Free to start, no credit card required.
Get Started Free →hashgraph-online/memory-load
Use when Wingman memory is already known enabled in the current repository and starting non-trivial coding, debugging, planning, review, API integration, or business-logic work, or when the user explicitly asks to load or read Wingman memory.
Passed 10 of 15 test cases with the skill, and 7 of 15 without it.
and 3 more in the full report
6 fixed9 passed either way6 still failing1 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 6 test cases the skill fixed. Same model, same prompt, with and without the skill.
The Domain Registry lists domains/auth.md with 'Related Domains: domains/audit.md'. My task only modifies password hash iteration counts in auth. Should I automatically read domains/audit.md?
| Checked | Without skill | With skill |
|---|---|---|
| Omits optional related domain reading when task does not touch that related scope | Missed | Passed |
The instructions your model reads when the skill loads.
memory-load is read-only. It loads current truth by default and history only when the task needs explanation or past-change context.
Use .wingman/memory/ as the repository-local memory root.
Repository memory is enabled only when both core entry files exist:
.wingman/memory/brief.md
.wingman/memory/context.md
If .wingman/memory/ is missing, continue without memory unless the user explicitly asked about memory. If the directory exists but a core entry file is missing, report partial memory only when the user asked about memory or the task depends on memory consistency.
Use this authority order when memory files disagree:
brief.md current Project Decisions
> domains/ current truths
> context.md hot working state
> history/ past events and evidence
History explains why something changed. It is not current truth unless a current entry in brief.md or domains/ links to it as evidence.
Only Status: current current-truth entries are binding. candidate, deprecated, and superseded entries are background only.
Relations:
updates <id> replaces the referenced entry for overlapping scope.extends <id> keeps both entries valid when compatible.derived_from <id-or-path> records evidence lineage, not authority.If two current entries with the same Subject conflict and no relation resolves them, stop and ask which rule is valid.
test cases passed, vs 7 / 15 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.