What is cli-continues security.instructions.md?

Security rules for a command-line tool that resumes sessions and passes options between coding tools. They cover safe process launching, permission flags, and the safety of generated handoff text.

It is for reviewing session resumption, argument forwarding, approval-mode precedence, and Markdown handoff output for security problems.

They reduce the risk that user-controlled session IDs or paths run unintended shell commands, and prevent automatic approval options from granting more access than requested.

cli-continues security.instructions.md — Instructions file by yigitkonur

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods

npx agentmods add instructions/yigitkonur/cli-continues/security

Clone the repo

git clone --depth 1 https://github.com/yigitkonur/cli-continues

Made for: GitHub Copilot.

Per session 419 This file is loaded in full into every session.

When invoked 419 The same file — it is already loaded in full.

Security scan A 0 findings. Scan, not verified.

Origin original No closer match found in the catalogue.

Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

Model	Per session	Once invoked
Fable 5	$0.00419	$0.00419
Opus 5	$0.00210	$0.00210
Sonnet 5	$0.00084	$0.00084
Haiku 4.5	$0.00042	$0.00042

Measured 2d ago against content hash 25ab7dc2f8df, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

cli-continues security.instructions.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.github/instructions/security.instructions.md · 36 lines

What it actually says

Security Review Guidelines — Core Utilities

Command Injection Prevention (resume.ts)

External processes must be spawned with spawn() and arguments as an array — NEVER exec() with string interpolation
Session IDs and file paths from parsed sessions are user-controlled and may contain shell metacharacters (;, |, &, $, backticks) — they must always be array elements, never embedded in a shell string

// Avoid — command injection if sessionId contains ; or | or $()
exec(`claude --resume ${sessionId}`);

// Prefer — safe regardless of sessionId content
spawn('claude', ['--resume', sessionId], { stdio: 'inherit' });

Forward Flag Security (forward-flags.ts)

--dangerously-skip-permissions (Claude) and --dangerously-bypass-approvals-and-sandbox (Codex) must ONLY be set when the source session explicitly requested auto-approve behavior
Flag precedence is security-critical: auto-approve > full-auto > sandbox > ask-for-approval — deviations from this order could grant unintended permissions in the target tool
Never map a "plan mode" flag or any scheduling flag to auto-approve behavior

Handoff Output Safety (markdown.ts)

Do NOT embed secrets, API keys, tokens, or environment variable values in handoff markdown output
Tool activity summaries (shell command output, file diffs) may contain sensitive data — the verbosity config caps limit exposure; do not bypass these caps
Home directory paths in handoff output must be tildified (~/) using safePath() — never expose the full absolute home path

General

No hardcoded credentials, API keys, or tokens anywhere in source files
User-supplied paths (session file paths, cwd values) must not be passed to shell execution without sanitization — use spawn() with array args only

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

2d ago First seen · 36 lines · 419 tokens per session scan A 25ab7dc2f8df

Subscribe to this mod's changes

cli-continues security.instructions.md is an instructions file published in the GitHub repository yigitkonur/cli-continues (1,479 stars, last pushed 3mo ago), licensed MIT. It adds 419 tokens to every session, about $0.0021 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Security ai-coding-assistant claude-code cli context-transfer cursor gemini-cli

skills CLAUDE.md

Instructions for lovstudio/skills, covering claude.md, what this is, repo layout, how users install and skills.yaml schema.

lovstudio/skills · 1,821 tokens

AgentHarbor CLAUDE.md

Instructions for abhiunix/AgentHarbor, covering claude.md, commands, architecture, two tauri windows and adapter trait — the core abstraction.

abhiunix/AgentHarbor · 2,538 tokens

AgentHarbor AGENTS.md

Instructions for abhiunix/AgentHarbor, a project described as: 🚢 One native app for all your AI coding agents - live rate limits, session usage, deep analysis of your & monthly spend for Claude Code, Cursor, Codex, Gemini etc plus one-click MCP/rules/skills/sub-agent deploy.

abhiunix/AgentHarbor · 206 tokens

cli-continues security.instructions.md