Index of every page on this site: llms.txt
All skills
open-neko avatar
by open-nekoopen-neko/openneko22 stars
3
Run it yourself

Model or interpret a generated CRM app, especially for Salesforce-shaped sales data. Use when a user wants a CRM, sales pipeline, account/contact/deal tracking, Salesforce replacement or import, opportunity stages, ownership mapping, or help interpreting a CRM import report.

  • 2 files
  • 4.9 KB
  • Updated 3 months ago
  • GitHub
Use this Skill: https://skilld.dev/gh/open-neko/openneko/crm

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

SKILL.md

≈70 tokens for metadata: the name and description. ≈1.1k when used: this file. ≈45 more on demand in 1 file.

Description uses 3.5% of example budget

Before choosing a Skill, your Agent reads its name and description. All available Skills share that space.

  • A shorter description leaves more room for other Skills. This entry exceeds our 1% size suggestion.

In our Claude Code example, all Skill names and descriptions share 8,000 characters. This Skill uses ≈281 characters, or 3.5%.

The 1% threshold is a size suggestion. Longer descriptions can still fit.

Your model, settings, and other Skills decide how much text your Agent can read.

Example settings and source

The example uses a 200k-token context and default Claude Code settings. The count includes the name, description, separators, and when_to_use when present. Codex also counts local file paths.

Skit's source and limits: Codex 0.160.1, Claude Code 2.1.292.

CRM

Apply CRM domain knowledge through the generic records engine. Use mcp_neko_records_browse_blueprints with blueprint: "crm" to load the canonical starting payload. Before proposing or changing the app, read ../app-builder/SKILL.md completely and follow it to adapt the loaded payload. Before reading or writing CRM rows, read ../records/SKILL.md completely. The blueprint is a prior, not a mandatory sales process.

Model the sales workflow

  • Account is the company or buying organization. Keep it org-visible unless the operator explicitly partitions accounts by owner.
  • Contact is a person related to an account. Preserve contacts without an account when legacy data permits them; report the missing relationship.
  • Opportunity is the deal or revenue event. Usually make it owner-visible and reference its account.
  • Activity is a call, meeting, email, note, or task. Model the user's actual follow-up workflow rather than copying every source activity subtype.
  • Custom Salesforce objects become ordinary record objects only when the operator needs their workflow. Never add CRM behavior to engine code.

Use reference fields for relationships, currency for monetary amount, date for close dates, and percent for probability. Keep source IDs as text during migration so relationships remain stable.

Stages and status

Ask for the pipeline stages when they materially differ from the canonical qualification → discovery → proposal → negotiation → closed-won/lost flow. Store durable snake_case values such as closed_won; labels may change later. Do not infer probability, forecast category, or a win/loss state from a stage name unless the operator defines that mapping.

For imported Salesforce data, preserve unfamiliar stage values and surface them for mapping. Never silently coerce an unknown value into the closest canonical stage.

Ownership

Words such as owner, rep, assignee, book of business, and "my deals" normally map to the records substrate's owner_user_id and object visibility: owner. Do not create a second ordinary owner field unless the source owner's raw ID must be retained for migration evidence.

An imported source owner is not automatically an OpenNeko user. Treat linked, unlinked, conflict, and ignored identity states distinctly. Never guess an owner match from a similar display name or email.

Salesforce-shaped imports

Map common source objects as follows, then adapt:

  • Account → account
  • Contact → contact, referencing account
  • Opportunity → opportunity, referencing account
  • Task and Event → activity when one combined timeline fits

Use source-scoped IDs and connector identity; two Salesforce instances may contain the same external ID. Preserve polymorphic WhoId and WhatId target sets rather than guessing one object. Do not manually rewrite 15-character Salesforce IDs; the connector's normalization owns that rule.

Baseline CSV import is insert-only. Interpret reports precisely:

  • inserted means a row committed through GraphJin;
  • duplicate means a stable key already existed and was not overwritten;
  • rejected means validation failed—summarize reasons and row locations;
  • unresolved reference or identity means the row may exist but needs mapping.

Do not claim a CSV load is a live mirror. Connected mirror mode remains read-only locally until an explicit verified cutover completes.

Optional starter watches

The canonical CRM blueprint includes three disabled-by-default starter_workflows. Offer them as explicit choices before proposing the app:

  • opportunities with no activity in 30 days;
  • records owned by an unlinked or departed source user;
  • deals closing this month, grouped by owner.

Preserve every choice in the app_create payload and enable only the watches the operator selected. Each enabled workflow uses a durable GraphJin watch as its primary wake-up path and a scheduled fallback, then publishes the same deterministic result into Briefing. App creation and terminal imports also produce Briefing summaries; use the attached import report for row, identity, and staging-cleanup facts rather than reconstructing them from chat history.

Working with CRM records

Use $records for every read or write. Resolve exact account, contact, opportunity, activity, and reference IDs before proposing changes. When several companies or people share a name, disambiguate with stable facts such as domain, email, city, account, stage, or close date.

Source: SKILL.md on GitHub

No rule matched.

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

No third-party reports yet.

Signed by skilld at 621f131. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 4 weeks ago.

Activeupdated 3 months ago
type
Skill

README badge

README badge for open-neko/openneko/crm