Claude-skill-registry activating-memories

Loads project-specific memories at session start and triggers onboarding for new projects. Use when starting a new coding session or when the user says "activate" or asks about project context.

install

source · Clone the upstream repo

git clone https://github.com/majiayu000/claude-skill-registry

Claude Code · Install into ~/.claude/skills/

T=$(mktemp -d) && git clone --depth=1 https://github.com/majiayu000/claude-skill-registry "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/data/activating-memories" ~/.claude/skills/majiayu000-claude-skill-registry-activating-memories && rm -rf "$T"

manifest: skills/data/activating-memories/SKILL.md

source content

<quick_start>

Check

./.{AGENT_NAME}/skills/memories/

for memory skill folders.

Each memory is a skill file at

./.{AGENT_NAME}/skills/memories/<name>/SKILL.md

Example baseline set:

```
project-overview/SKILL.md
```
```
suggested-commands/SKILL.md
```
```
style-and-conventions/SKILL.md
```
```
task-completion-checklist/SKILL.md
```

If empty or missing, read
```
references/onboarding-guide.md
```
and execute onboarding.
Otherwise, read memories and proceed.
If you learn new info, update the relevant
```
<name>/SKILL.md
```
via
```
$writing-memories
```
. </quick_start>

<decision_points>

Memories missing/empty -> read
```
references/onboarding-guide.md
```
and run onboarding. </decision_points>

<quality_checklist>

Each memory should be dense with high-signal, repo-specific facts.
Prefer concrete evidence (paths, filenames, commands) over vague summaries.
If a memory is sparse, expand it by scanning the repo and updating it. </quality_checklist>

<failure_modes>

```
AGENT_NAME
```
unclear: list directories under
```
.
```
and choose the active one (e.g.
```
.codex/
```
or
```
.claude/
```
). </failure_modes>