AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
SKILL verified MIT Self-run

Deepseek Cli

skill-appleweiping-weiping-whale-deepseek-cli · by appleweiping

A Claude skill from appleweiping/WEIPING_WHALE.

No reviews yet
0 installs
5 views
0.0% view→install

Install

$ agentstack add skill-appleweiping-weiping-whale-deepseek-cli

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No 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.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/skill-appleweiping-weiping-whale-deepseek-cli)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
2mo ago

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Deepseek Cli? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

DeepSeek CLI Skill

Use this skill for non-trivial work on deepseek-cli-agent, especially release-level improvements, runtime honesty, safety, MCP, memory, terminal UX, packaging, or documentation.

Purpose

Keep DeepSeek CLI honest, small, useful, and release-ready. The project is a terminal-native coding agent, not a web app and not a toy demo. Every substantial change should improve a complete lifecycle:

planning -> design -> development -> testing -> release -> maintenance -> monitoring -> upgrade iteration

Phase 1 - Plan

  1. Inspect the current git status and protect user changes.
  2. Read README.md, package.json, config.toml, src/index.ts, src/agent.ts, src/config.ts, src/llm/deepseek.ts, src/memory.ts, src/session.ts, src/safety/*, src/tools/*, src/mcp/*, .github/workflows/ci.yml, and the relevant tests/scripts.
  3. Identify the project relationship:
  • vipin-council: orchestration and cross-model review.
  • deepseek-cli: terminal worker and fast local coding agent.
  • vipin-lab: experiment/workbench surface.
  1. Define a version-level batch. Do not ship only typo fixes unless the user explicitly asks for a tiny change.

Phase 2 - Design

For non-trivial upgrades, inspect concrete source files from active open-source agent CLIs. Do not rely only on READMEs.

Recommended references:

  • Google Gemini CLI: config validation, storage/session paths, retry/error handling, tool execution tests.
  • Aider: command running, model metadata, IO and regression tests.
  • OpenCode: config schema, permission model, provider status, session and error structure.

Record the specific modules inspected in the README, release notes, or handoff when they shape the change.

Phase 3 - Develop

Prefer small internal modules with clear responsibilities:

  • runtime version constants
  • safe error redaction
  • config diagnostics
  • session storage
  • memory outbox
  • MCP diagnostics
  • safety gates

Rules:

  • No raw API keys, bearer tokens, provider response bodies, or private paths in user-facing diagnostics.
  • Do not depend on retired Agent Hub mailbox/state paths.
  • Use active agentmemory when reachable; otherwise write only to the local DeepSeek CLI outbox.
  • Keep the default permission model safe: preview writes, workspace sandbox, shell approvals on request.
  • Keep runtime claims measurable. If the bundle size changes, remove stale size slogans.

Phase 4 - Test

Before review, run the full local gate:

npm run typecheck
npm test
npm pack --dry-run

npm test must build, run no-network E2E checks, and run the release scan.

For changes touching API, memory, sessions, safety, MCP, or config, add or update E2E coverage. Test offline/failed-service behavior, not only the happy path.

Phase 5 - Release

  1. Update package.json, runtime version constants, README, and package-lock when the change is version-level.
  2. Ensure config.toml is safe as a packaged fallback.
  3. Verify npm pack --dry-run contains only intended release files.
  4. After verification and review, commit and push to the repo remote.

Phase 6 - Maintenance

Keep these surfaces current:

  • README.md
  • .github/workflows/ci.yml
  • scripts/e2e.mjs
  • scripts/release-scan.mjs
  • config.toml
  • this skill file

Remove stale claims instead of layering corrections underneath them.

Phase 7 - Monitoring

Use these checks when diagnosing user reports:

deepseek --doctor --json
deepseek --models --json
npm test
npm pack --dry-run

Inspect whether failures are in auth, endpoint config, MCP startup, session loading, memory save, shell approval, or file sandboxing.

Phase 8 - Upgrade Iteration

For each substantial upgrade:

  1. Re-read this skill and live project files.
  2. Inspect at least one concrete source module from a mature open-source CLI relevant to the upgrade.
  3. Implement the full batch.
  4. Run all verification.
  5. Only after the batch is complete, ask parallel reviewers to score architecture, safety, UX, tests, documentation, release readiness, and honesty.
  6. Iterate until the score is 10/10 before commit and push.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.