CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

memory-architecture

Use when the user asks how you work (memory, recall, dreams, curiosity, hormones, how you learn skills, what is switched on), or when you need to interpret memory_status or self_state output. Not for recalling facts (use memory_search) or for saving notes (use working_memory_note).

Quality

78%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./skills/memory-architecture/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

75%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A well-structured, information-dense reference skill with genuinely actionable rules (self_state verification, honest uncertainty, explicit tool routing) and no filler. Its gaps are minor: some redundancy around the self_state rule, descriptive sections without corresponding guidance, and a buried pointer to deeper docs.

Suggestions

State the "call self_state before claiming a part is running" rule once (in "Talking about yourself") and reference it elsewhere instead of repeating it in the intro and owner-controls sections.

Move the pointer to the detailed docs under `memory/` out of the inline sentence and into a clearly labeled final section (e.g., "## Deeper detail") so navigation to it is explicit.

DimensionReasoningScore

Conciseness

The body is dense and assumes competence — it documents bespoke system behavior Claude could not know (lifecycles, hormone half-lives, RRF fusion) with no padding about generic concepts. It falls short of anchor 5 because of minor redundancy: the "call self_state before saying a part is running" rule appears in the intro, again under "Talking about yourself", and indirectly under "The owner's controls", and several sentences carry multiple parentheticals that could be tightened.

4 / 5

Actionability

Concrete directives dominate: "Call `self_state` before saying a part is running", "say what you searched", "read this file before answering", "do not promise a deletion you did not do", and a complete tool-routing table in "When to use what" (e.g., "Prior work, decisions, people, dates: `memory_search`, then `memory_get`"). Not a 5 because sections like Feelings and Dreams are largely descriptive background with little guidance on what to do with the information beyond tone effects.

4 / 5

Workflow Clarity

The core workflows are clearly sequenced with checkpoints: answer questions about how parts work from this file, verify live state via `self_state` before claiming a part runs, and fall back to "say you are not sure instead of filling the gap" — an explicit error-recovery guardrail. It is not a 5 because some sequences (e.g., handling a correction via reconsolidation vs. memory_pin) are described but not stepped through, and no explicit validate-then-retry loop exists for the owner-requested forget/edit path.

4 / 5

Progressive Disclosure

No bundle files exist, and the single ~70-line body is well-organized into clearly headed, navigable sections that appropriately cover one cohesive topic. It fits anchor 4's good structure with minor gaps rather than 5: it exceeds the under-50-lines simple-skill case, and the pointer to further detail ("the docs under `memory/` have the detail") is buried inline in "Talking about yourself" rather than clearly signaled as a navigable reference section.

4 / 5

Total

16

/

20

Passed

Description

73%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A well-crafted trigger-oriented description with explicit use-when guidance, specific tool names, and excellent negative-space disambiguation. Its main weaknesses are an implicit rather than stated "what", second-person voice, and a few missing natural trigger synonyms.

Suggestions

State the capability explicitly in third person, e.g., "Explains how this assistant's memory, recall, dreams, curiosity and skill learning work. Use when the user asks how you work...".

Add natural trigger synonyms users would say, such as "feelings", "mood", "forgetting", or "what do you know about me".

DimensionReasoningScore

Specificity

The description enumerates concrete capability areas ("memory, recall, dreams, curiosity, hormones, how you learn skills, what is switched on") and one concrete action ("interpret memory_status or self_state output"), which initially matches anchor 4's several specific items. However, it uses second-person voice ("how you work", "when you need to interpret"), which the guidelines penalize by reducing specificity by 1, landing at 3. It is not a 2 because it goes far beyond naming a bare domain — it names specific subsystems and two named tool outputs.

3 / 5

Completeness

The "when" is explicit and concrete ("Use when the user asks how you work (...), or when you need to interpret memory_status or self_state output") with a clear negative boundary. The "what" is present but implicit — the skill's explanatory purpose must be inferred from the trigger phrasing rather than stated as a capability (e.g., "Explains how your memory, recall and dreams work"), which fits anchor 4 rather than anchor 5's fully explicit both-what-and-when.

4 / 5

Trigger Term Quality

Strong natural phrases a user would actually say: "how you work", "memory", "dreams", "curiosity", "hormones", "how you learn skills", "what is switched on", plus exact tool names (memory_status, self_state). It falls short of anchor 5's comprehensive synonym coverage — natural terms like "feelings", "mood", "forgetting", or "what do you know about me" (which the body itself routes to this skill's tools) are absent from the description.

4 / 5

Distinctiveness Conflict Risk

It occupies a clear niche (self-explanation) with distinct triggers, and explicitly disambiguates against the nearest competing functions: "Not for recalling facts (use memory_search) or for saving notes (use working_memory_note)". This is stronger than anchor 4's minor overlap risk — the exclusions make wrong-skill triggering unlikely.

5 / 5

Total

16

/

20

Passed

Validation

87%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 14 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

metadata_version

'metadata.version' is missing

Warning

metadata_field

'metadata' should map string keys to string values

Warning

Total

14

/

16

Passed

Repository

Bitterbot-AI/bitterbot-desktop

Path

skills/memory-architecture/SKILL.md

Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.