Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Memory is the cornerstone of intelligent agents. Without it, every interaction starts from zero. This skill covers the architecture of agent memory: short-term (context window), long-term (vector stores), and the cognitive architectures that organize them.
.claude/skills/sickn33-agent-memory-systems/SKILL.md| Model | Eval pass | Runs |
|---|---|---|
| gemini-3.6-flash | 100% | 61 |
| gemini-3.1-pro-preview | 100% | 1 |
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✗→✓ | ▲ Improved | 309% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 270% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 238% | 0% |
| case-20 | ✗→✓ | ▲ Improved | 206% | 0% |
| case-22 | ✗→✓ | ▲ Improved | 536% | 0% |
Memory is the cornerstone of intelligent agents. Without it, every interaction starts from zero. This skill covers the architecture of agent memory: short-term (context window), long-term (vector stores), and the cognitive architectures that organize them.
Key insight: Memory isn't just storage - it's retrieval. A million stored facts mean nothing if you can't find the right one. Chunking, embedding, and retrieval strategies determine whether your agent remembers or forgets.
The field is fragmented with inconsistent terminology. We use the CoALA cognitive architecture framework: semantic memory (facts), episodic memory (experiences), and procedural memory (how-to knowledge).
Read the detailed guide before executing this skill. It retains the complete procedure and reference material. Treat its safety, prerequisites, and validation requirements as mandatory. For focused work, load the relevant sections; for end-to-end work, read the guide completely.
User request:
> Use @agent-memory-systems for this task: Memory is the cornerstone of intelligent agents.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 24,046 | 16,502 | -31% | 1 | 1 | 0% | 5,531 | 11,685 | +111% | 0 | 0 | — |
case-02 | fail→fail | 16,208 | 18,044 | +11% | 1 | 1 | 0% | 3,676 | 11,406 | +210% | 0 | 0 | — |
case-03 | fail→fail | 33,909 | 23,725 | -30% | 1 | 1 | 0% | 6,194 | 12,444 | +101% | 0 | 0 | — |
case-04 | fail→fail | 24,770 | 17,914 | -28% | 1 | 1 | 0% | 4,183 | 10,735 | +157% | 0 | 0 | — |
case-05 | fail→pass | 15,068 | 12,051 | -20% | 1 | 1 | 0% | 2,428 | 9,926 | +309% | 0 | 0 | — |
case-06 | fail→pass | 16,700 | 12,213 | -27% | 1 | 1 | 0% | 2,685 | 9,944 | +270% | 0 | 0 | — |
case-07 | pass→pass | 7,419 | 5,217 | -30% | 1 | 1 | 0% | 1,336 | 8,779 | +557% | 0 | 0 | — |
case-08 | pass→pass | 16,927 | 12,394 | -27% | 1 | 1 | 0% | 2,756 | 9,835 | +257% | 0 | 0 | — |
case-09 | pass→pass | 14,136 | 12,295 | -13% | 1 | 1 | 0% | 2,445 | 10,203 | +317% | 0 | 0 | — |
case-10 | pass→pass | 18,818 | 14,117 | -25% | 1 | 1 | 0% | 3,023 | 10,602 | +251% | 0 | 0 | — |
case-11 | pass→pass | 12,954 | 12,458 | -4% | 1 | 1 | 0% | 2,872 | 10,272 | +258% | 0 | 0 | — |
case-12 | fail→pass | 17,246 | 11,814 | -31% | 1 | 1 | 0% | 2,996 | 10,125 | +238% | 0 | 0 | — |
case-13 | pass→pass | 10,521 | 10,191 | -3% | 1 | 1 | 0% | 1,876 | 9,734 | +419% | 0 | 0 | — |
case-14 | pass→pass | 24,477 | 15,793 | -35% | 1 | 1 | 0% | 4,200 | 10,979 | +161% | 0 | 0 | — |
case-15 | pass→pass | 14,036 | 8,227 | -41% | 1 | 1 | 0% | 2,489 | 9,388 | +277% | 0 | 0 | — |
case-16 | pass→pass | 13,373 | 8,721 | -35% | 1 | 1 | 0% | 2,266 | 9,316 | +311% | 0 | 0 | — |
case-17 | pass→pass | 12,496 | 10,090 | -19% | 1 | 1 | 0% | 2,136 | 9,651 | +352% | 0 | 0 | — |
case-18 | pass→pass | 13,716 | 9,423 | -31% | 1 | 1 | 0% | 2,355 | 9,528 | +305% | 0 | 0 | — |
case-19 | pass→pass | 15,939 | 11,906 | -25% | 1 | 1 | 0% | 3,090 | 10,187 | +230% | 0 | 0 | — |
case-20 | fail→pass | 14,931 | 4,397 | -71% | 1 | 1 | 0% | 2,841 | 8,701 | +206% | 0 | 0 | — |
case-21 | pass→pass | 6,904 | 5,717 | -17% | 1 | 1 | 0% | 1,190 | 8,801 | +640% | 0 | 0 | — |
case-22 | fail→pass | 8,502 | 5,687 | -33% | 1 | 1 | 0% | 1,385 | 8,814 | +536% | 0 | 0 | — |
case-23 | pass→pass | 20,222 | 12,670 | -37% | 1 | 1 | 0% | 2,227 | 10,253 | +360% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted. The headline lift of +22 percentage points is the difference between those two pass rates over the 23 comparable cases.
The publisher has shipped newer versions since this run, so these numbers describe v1, not the version currently listed.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.