aidc provisions and runs container images that hold API tokens (Anthropic, Z.ai, OpenRouter, self-hosted Anthropic-compatible servers) and execute AI coding agents with workspace and network access. Relevant attack surfaces:

• the host-side launcher (bin/aidc, lib/aidc.sh, install.sh) and the templates it writes into a user's repo (templates/)
• the host clipboard bridge (bin/aidc-clipboard-server) and its Unix-socket protocol
• container hardening defaults (egress firewall, supply-chain shims, scanner pre-installs)
• the Claude-profile alias generation in ~/.local/bin

Bugs in claude, codex, opencode, cursor-agent, pmg, vet, rtk, or the upstream devcontainer base image are out of scope — report those to their respective projects.

aidc is pre-1.0 and rolling-release. Only main is supported; fixes land there. Tagged releases are best-effort.

Do not open a public issue for security bugs.

Preferred: open a private GitHub Security Advisory at https://github.com/cyfinoid/aidc/security/advisories/new.

Alternative: email security@cyfinoid.com with [aidc] in the subject. Include:

• a description of the issue and impact
• reproduction steps or a proof-of-concept
• the commit hash you tested against
• any suggested mitigation

You should get an acknowledgement within 5 business days. Coordinated disclosure timeline is negotiable; default is 90 days from acknowledgement.

• Findings that require an attacker to already have root on the host or full write access to the user's home directory.
• Token leakage caused by the user committing .env files into their own repo (the scaffold ignores them; user override is user-owned risk).
• Denial of service against the local Docker daemon.
• Issues only reproducible against unsupported architectures or non-macOS hosts.