القائمة
Add a memory system to an agent — file awareness, knowledge graph, structured state, or multi-session tracking
يُحدَّد Prompt الذي يراجع المصدر أولًا بشكل افتراضي. يمكنك التبديل إلى أمر مباشر أو تنزيل نسخة محلية.
اقرأ SKILL.md وأي ملفات مرافقة يعرضها SkillsMP قبل أن تقرر التثبيت.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
يتجاوز الأمر المباشر Prompt المخصّص للمراجعة. افحص المصدر قبل تشغيله.
npx skills add https://github.com/Abilityai/trinity-skills --skill add-memoryيبقى الأمر في سطر واحد. مرّر أفقيًا لمراجعته كاملًا قبل النسخ.
تفضّل نسخة محلية؟ نزّل الملفات المتاحة حاليًا لدى SkillsMP.
Explain a complex topic back to you with visuals - the right diagram for each facet (flowchart, timeline, sequence, schema/ER, state machine, chart, comparison, mind map) plus optional conceptual illustrations - rendered in a style you control, then bundled onto a single page. SELF-CONTAINED: it explains what was already discussed or provided; it does NOT research. Decomposes the topic into the facets that most need to be SEEN, picks a visual modality per facet, renders structural diagrams via themed Mermaid (HTML->PNG, $0) and conceptual facets via /create-explanatory-image, then composes /one-pager to assemble everything into one dense page + PDF. Pick the visual style with --style (brand-dark default, clean-light, blueprint, whiteboard, custom). Use when someone says "explain this to me visually", "show me how this works", "diagram what we just discussed", "make this make sense", or wants a visual brief of a topic.
Generate a brand-locked, information-dense single-page PDF that lets a reader understand an entire topic - a company, a product, a project, a status update, any subject - in one page. Distills a raw brief into a composed grid of reusable blocks (hero-stat, KPI strip, two-column prose, comparison table, timeline, callout, icon-grid, fact-rail, mini-chart), renders one HTML via Playwright to a single PNG, mechanically guarantees it fits on ONE page via a fit-guard + tightening ladder, then wraps it as a 1-page PDF. Default 16:9 landscape; A4/Letter portrait via flag. Dark brand canvas default; light print-friendly theme via flag. Use when someone needs a dense one-pager / one-sheet / fact sheet / briefing doc that ships as a PDF.
Generate a brand-locked multi-slide presentation deck (16:9, 1920x1080) and assemble into a PDF. Mirrors the social-carousel pattern - each slide is an HTML+Playwright PNG render with its own textless soft-futurism hero image. Final step combines all slide PNGs into a single PDF for distribution. Six slide types - cover, section, content, statement, closing on the shared dark canvas brand system, plus a dense type composed via /one-pager for information-dense slides (dashboards, fact sheets, at-a-glance briefings). Use when building a workshop deck, conference talk, internal presentation, or any 16:9 slide set that needs to look brand-coherent and ship as a PDF.
عرض SKILL.md
| name | add-memory |
| description | Add a memory system to an agent — file awareness, knowledge graph, structured state, or multi-session tracking |
| argument-hint | [memory-type] |
| disable-model-invocation | false |
| user-invocable | true |
| allowed-tools | Read, Write, Bash, Glob, Grep, AskUserQuestion |
| metadata | {"mirror":"abilities@ddf0420 plugins/agent-dev/skills/add-memory","version":"1.1","created":"2026-04-16T00:00:00.000Z","updated":"2026-05-16T00:00:00.000Z","author":"Ability.ai","changelog":["1.1: Add portability note — memory skills are copied into the agent so it stays self-contained with no plugin dependency","1.0: Initial version — adds a memory system (file-index, brain, json-state, or workspace) to an existing agent"]} |
| category | agent-development |
ℹ️ First, set expectations: before anything else, print one short line with this skill's version and its most recent change — the top entry of
metadata.changelogabove — e.g.add-memory vX.Y — recent: <summary>. Then proceed.
Add a memory system to an existing agent. Memory skills are COPIED into the agent's .claude/skills/ directory — the agent becomes self-contained with no plugin dependency.
| Type | Description | Skills Installed |
|---|---|---|
| file-index | File system awareness — creates and maintains a searchable file index | setup-index, refresh-index, search-files |
| brain | Zettelkasten-style knowledge graph — atomic notes with connections | setup-brain, create-note, search-brain, find-connections |
| json-state | Structured JSON state with jq-based updates and action logging | setup-memory, load-memory, update-memory, memory-jq |
| workspace | Multi-session project tracking — organized folders, sessions, archives | setup-projects, create-project, create-session, archive-project |
Check if we're in an agent directory (has CLAUDE.md) or if user specified a path.
# Check current directory
if [ -f "CLAUDE.md" ]; then
AGENT_DIR="$(pwd)"
else
# Ask user for agent path
fi
If no CLAUDE.md found, ask:
If user provided an argument (e.g., /agent-dev:add-memory brain), use that.
Otherwise, use AskUserQuestion:
Question: "What kind of memory does this agent need?"
Header: "Add Memory"
Options:
File awareness (file-index) — "I need the agent aware of workspace files"
Knowledge graph (brain) — "I need connected notes, evolving knowledge"
Structured state (json-state) — "I need counters, config, structured data"
Multi-session workspace (workspace) — "I need organized projects and sessions"
Memory templates are stored in this plugin at:
agent-dev/memory-templates/
file-index/ # setup-index/, refresh-index/, search-files/
brain/ # setup-brain/, create-note/, search-brain/, find-connections/
json-state/ # setup-memory/, load-memory/, update-memory/, memory-jq/
workspace/ # setup-projects/, create-project/, create-session/, archive-project/
Copy the selected memory type's skills into the agent:
# Get the plugin directory (where this skill lives)
PLUGIN_DIR="$(dirname "$(dirname "$(dirname "$0")")")"
TEMPLATES_DIR="$PLUGIN_DIR/memory-templates"
# Copy skills to agent
cp -r "$TEMPLATES_DIR/[selected-type]/"* "$AGENT_DIR/.claude/skills/"
Add a section documenting the new memory capabilities:
## Memory: [Type Name]
This agent has [type] memory installed. Available skills:
- `/[skill-1]` — description
- `/[skill-2]` — description
...
Run `/[setup-skill]` to initialize.
Output:
## Memory Added
Installed **[type]** memory to `[agent-name]`.
### New Skills
- `/[skill-1]` — description
- `/[skill-2]` — description
### Next Step
Run `/[setup-skill]` to initialize the memory system.
~/. If using the native Claude Code memory: frontmatter field on custom subagent definitions, use memory: project (stores in .claude/agent-memory/, stays in repo) — never memory: user (stores in ~/.claude/, breaks portability)