Context
The skill is script-heavy and evolving fast. When Python dependencies are missing, agents currently discover failures late (mid-command) and then waste tokens installing packages ad-hoc. This also reduces reproducibility across machines.
Goal
Provide a one-shot, stdlib-only bootstrap script that:
- creates a local venv,
- installs the self-contained skill package (editable),
- optionally installs heavier indexing/embedding dependencies.
Non-Goals
- Do not implement a server runtime.
- Do not require users to maintain a separate
requirements.txt.
- Do not install anything globally; keep installs within a repo-local venv.
Approach
- Add
skills/kano-agent-backlog-skill/scripts/dev/install_prereqs.py (stdlib-only).
- Default behavior: create
.venv/ and install skills/kano-agent-backlog-skill with [dev] extras.
- Optional flag:
--with-embeddings to install embedding/FAISS deps (best-effort).
- Document the entrypoint in repo
README.md and AGENTS.md.
Alternatives
- Let scripts fail and ask the agent to install missing deps (current; wastes tokens).
- Commit a frozen lockfile and require exact sync (too heavy for pre-alpha demo).
Acceptance Criteria
- A single command exists that creates a venv and installs required deps for the skill scripts.
- The command is documented in
README.md and AGENTS.md.
- The script uses only stdlib so it can run before any dependencies are installed.
Risks / Dependencies
- Some embedding deps are platform-dependent; keep them optional and best-effort.
Worklog
2026-01-10 02:05 [agent=codex] Planned: add a one-shot script to create venv and install required Python deps so agents don't waste tokens installing ad-hoc.
2026-01-10 02:11 [agent=codex] Ready: bootstrap script and docs are specified with acceptance criteria.
2026-01-10 02:11 [agent=codex] Done: added stdlib-only install script and documented usage in README.md and AGENTS.md.