Use when the friend wants to see what the system remembers about a project (or the wiki overall) — "what do you remember about X", "show me my project memory." Triggers on the /ren:remember [project] slash command. Renders the L2 map in human prose, not a raw markdown dump. The cheapest trust-builder in the scope (spec §3.8 A-10) — read-only, always.
Scanned 9/2/2026
Install to Claude Code
npx -y skills add hazarsozer/ren-os --skill remember --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Remember?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/hazarsozer-remember)More formats (shields.io, HTML) on the badges page.
---
name: remember
description: |
Use when the friend wants to see what the system remembers about a
project (or the wiki overall) — "what do you remember about X", "show me
my project memory." Triggers on the /ren:remember [project] slash command.
Renders the L2 map in human prose, not a raw markdown dump. The cheapest
trust-builder in the scope (spec §3.8 A-10) — read-only, always.
version: 0.8.5
license: MIT
framework_version: "0.8.5"
schema_version: 1
type: skill
execution_tier: deterministic
contract:
required_outputs:
- "Human-readable rendering of the resolved L2 map, or a friendly no-memory-yet message — never a crash"
budgets:
turns: 1
files_written: 0
duration_seconds: 5
permissions:
read:
- "~/.renos/wiki/**"
write: []
execute: []
completion_conditions:
- "A rendering is printed to the user; no file is written or modified"
output_paths: []
tags: [read-only, trust-builder, l2-map, remember]
related_skills: [recall, ingest-project, bootstrap-project]
references_required: []
references_on_demand: []
---
# remember
"Show me what you remember about this project" — spec §3.8 A-10's cheapest trust-builder. **Read-only, always**: this skill never proposes, approves, applies, or otherwise touches the write-queue. It only reads an already-written L2 map and renders it as prose a friend can actually read, instead of a raw markdown dump of frontmatter and bullet-pointer syntax.
## When to use this skill
- Friend invokes `/ren:remember [project-slug]` — with a slug, the named project's map; without one, the master wiki index.
- Friend asks in plain language: "what do you remember about `<project>`", "show me my project memory", "what do you know about this wiki."
## When NOT to use this skill
- Friend wants to look up something SPECIFIC (a decision, a lesson) → `/ren:recall <query>`, not this — `/ren:remember` shows the whole map, `/ren:recall` searches.
- Friend wants to CHANGE what's remembered → `/ren:pin`, not this.
## Behavior
1. Call `skills.remember.lib.remember(project_slug)`.
2. Print the result verbatim. It already handles every case:
- a real map exists → title line, `Knowledge:`/`Decision map:`/`Recent log:` sections in prose (decision pointers drop their `(write_id)` — that's provenance plumbing, not something a friend needs to see), and a footer with counts.
- the map is quarantined (unreviewed, auto-generated by `/ren:ingest-project`) → a prominent warning line, before anything else.
- an unknown project slug → a friendly one-line pointer to `/ren:ingest-project`/`/ren:bootstrap-project`.
- no slug and nothing to show at all → what IS known (identity set up? how many global pages?) plus how to get started.
Never raises. There is no failure mode here that should ever look like a crash to the friend — worst case is "I don't have memory for that yet," never a traceback.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!