agentscope-ai / agentscope-ai/QwenPaw

[Feature]: Commands that are misspelled in the message channel are still forwarded to the agent

Aperta
#7,479 2 commenti 0 reazioni 0 assegnatari Vedi su GitHub
enhancement
Lingua principale
Python
Stelle
34.9k
Fork
3.1k
Merge medio
1g 15h
PR unite (30g)
225

Descrizione

## QwenPaw Version

2.0.0-beta.5

## Description

When I send incorrect commands to the QQ(or others) channel, the agent can still receive these commands and process them as messages instead of discard these erroneous commands and inform the user.

For example, spelling `/new` as `/mew`. At this point, the model will unexpectedly start working.

In addition, the `/help` command is not currently supported, which makes it difficult for users to remember which commands are available when they forget them, and the specific command format.

## Component(s) Affected

- [x] Core / Backend (app, agents, config, providers, utils, local_models)
- [ ] Console (frontend web UI)
- [x] Channels (DingTalk, Feishu, QQ, Discord, iMessage, etc.)
- [ ] Skills
- [ ] CLI
- [ ] Documentation (website)
- [ ] Tests
- [ ] CI/CD
- [ ] Scripts / Deploy

## Actual vs Expected

- **Actual:** These erroneous commands will be passed directly to Agent.
- **Expected:**
1. The command should be rejected, and the user should be told that the command is incorrect and that they should enter the `/help` command to view relevant help.
2. If the user is using a skill, which skill name should be provided that is closest to the corresponding skill

Guida per i contributori

Apri la guida per i contributori

Valutazione

Questa issue non è ancora stata valutata.

Ricevi le nuove issue nella tua casella

Un breve riepilogo di issue GitHub adatte ai principianti.