"Write, propose, commit, repair, and nudge Potpie Context Graph
Scanned 9/8/2026
Install to Claude Code
npx -y skills add VectorSpaceLab/AREX-Skill --skill graph-write --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Graph Write?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/vectorspacelab-graph-write)More formats (shields.io, HTML) on the badges page.
---
name: graph-write
description: "Write, propose, commit, repair, and nudge Potpie Context Graph
records safely."
metadata:
disco-role: operating
disable-model-invocation: true
license: Apache 2.0
---
# Potpie graph write
Use this sub-skill when the task is about adding durable memory, proposing graph mutations, committing plans, handling inbox items, running graph quality checks, or nudging agents from graph evidence.
## Read this when
- The user asks about `potpie record`, `potpie graph propose`, `potpie graph commit`, `potpie graph mutate`, `potpie graph mutation-template`, `potpie graph bulk apply`, `potpie graph import`, `potpie graph repair`, `potpie graph history`, `potpie graph inbox`, `potpie graph quality`, or `potpie graph nudge`.
- A write fails because a mutation payload is malformed, a plan needs review, a plan expired, or verification did not pass.
- You need to distinguish the canonical propose/commit flow from the legacy `graph mutate` wrapper.
## Do not use this for
- Discovery-only reads: read `../graph-read/SKILL.md`.
- Pot/source setup or default binding: read `../workspace-boundaries/SKILL.md`.
- Runtime/daemon startup: read `../runtime/SKILL.md`.
- Credential acquisition: read `../auth-integrations/SKILL.md`.
## Operating procedure
1. Prefer the canonical plan door: `graph propose` to validate and stage, then `graph commit` with explicit approval and verification as needed.
2. Use `graph mutation-template` before authoring payloads when the DSL or entity label is uncertain.
3. Keep mutations flat and retrieval-grade: durable descriptions, explicit scopes, source references, and no nested legacy wrappers unless the CLI specifically emits them.
4. Use `graph inbox`, `graph quality`, `graph history`, and `graph inspect` to review pending, stale, or questionable graph state before applying more writes.
5. Treat `record` as a convenient durable-memory write path and `graph nudge` as a bridge from graph evidence to agent guidance.
## References
- `references/workflow.md` — write command matrix, flat mutation DSL, approval/verification, inbox, quality, bulk, and nudge flows.
- `references/troubleshooting.md` — malformed payloads, review gates, expired plans, bulk failures, inbox states, and quality repairs.
## Verification notes
- Safe native candidates include graph CLI contract, graph workbench plan/inbox/quality tests, nudge service, apply-once, semantic mutation validation, semantic mutations, graph-plan compatibility, and graph-workbench ontology tests.
- No accelerator backend is required for write-side CLI/contract verification.
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!