Install
$ agentstack add skill-alfredxw-nova-teller-config ✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
About
Teller Config
Use this skill before calling write_tellers.
Workflow
- Call
list_tellersfirst. For updates, callread_tellersfor the exact teller IDs. - Use
write_tellersfor create/update/delete. Do not edit teller JSON files directly. - Built-in tellers can be read and copied as examples. Deleting built-in tellers is rejected.
- For update, preserve slots and policy fields the user did not ask to change.
- For delete, require an explicit user request.
Teller Shape
Important fields:
id: stable ID. Required for update/delete; create may generate one if omitted.name: user visible name.description: short explanation of the teller's narrative role.random_event_rate: number between 0 and 1. Use small values unless the user wants surprise events.tags: short searchable labels.context_policy: controls which context groups the teller expects.slots: prompt slots used by the interactive story agent.
Do not change version, path, custom, invalid, error, created_at, or updated_at unless preserving an existing complete object from read_tellers.
Context Policy
context_policy contains:
creator: how to use CREATOR.md and creator-level rules.lore: how to use lore/context library.runtime_state: how to use current story state and turn context.
Keep these as short policy strings. They guide prompt assembly but do not replace runtime safety rules.
Slots
Each slot contains:
id: stable slot ID.name: user visible slot name.target: where the slot applies, such assystem,turn_context, or another existing target read from a teller.enabled: boolean.content: prompt text for that target.
When modifying slots:
- Preserve slot IDs so existing UI selection and semantics remain stable.
- Keep slot content focused on narrative behavior, not backend tool permissions.
- Do not put story facts, chapter prose, or temporary scene state into teller slots.
- If a new slot target is needed, mirror the target style already present in existing tellers.
Style Rules
style_rules maps scenes to inline style reference content:
scene: scenario label.style_contents: list of text snippets used as prose style references. Each item is stored as text, not a file path, and should stay within 8000 characters.
Only add style rules when the user asks for scene-specific style behavior or when an existing teller already uses that pattern.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: alfredxw
- Source: alfredxw/denova
- License: Apache-2.0
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet — be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.