skills.rest
Arasz
AraszCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

ai-badger-project-ops

Guides ai-badger framework refresh, skill enablement, and Hermes symlink maintenance across projects.

2|Updated Aug 2, 2026
Search Tags:#scaffolding#skill management#symlinks#hermes#git staging#ai-badger#framework refresh
One-click install
npx skills add https://github.com/Arasz/ai-raccoon --skill ai-badger-project-ops-arasz
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: ai-badger-project-ops
Source: https://github.com/Arasz/ai-raccoon/tree/main/.ai-badger/skills/learned/uncategorized/ai-badger-project-ops
Command: npx skills add https://github.com/Arasz/ai-raccoon --skill ai-badger-project-ops-arasz

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Running ai-badger framework operations (scaffolding, den-refresh version bumps, opt-in skill enablement) involves non-obvious pitfalls: Hermes skill symlinks break when scaffolding runs in a task worktree, refresh JSON reports are misread, staging sweeps in foreign session state, and pre-commit gates fail on unstaged scaffold outputs. This Skill captures the verified operational knowledge to run these workflows correctly. ## Core Features & Use Cases - Hermes skill-link lifetime rules: Explains that ~/.hermes/skills/<project>/* are symlinks into the scaffold target, why scaffolding must run on the durable checkout, and how relink_hermes_skills repairs dangling links. - Refresh runbook: Covers locating the framework checkout when $AI_BADGER is unset, using the framework venv python, capturing the first-run JSON report, staging only the refresh footprint, and handling git rebase --autostash under concurrent sessions. - Opt-in skill enablement: Shows the include.skills config.json shape, the documentation group expansion, and idempotency verification via git status and link counts. - Use Case: When asked to bump a project from ai-badger 0.79.0 to 0.80.0, follow the worked examples to run refresh.py with explicit --root and --target, exclude foreign state.json dirt, commit with the repo's convention, and push to a green merge. ## Quick Start Ask the agent to refresh this project's ai-badger framework to the latest version and commit the result following the repo's history convention.

Frequently Asked Questions about ai-badger-project-ops

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run an ai-badger den-refresh on a project?▼

Locate the framework checkout, run refresh.py with the framework venv python passing explicit --target and --root, and redirect the full JSON report to a file. The first run performs the re-scaffold; a second run confirms convergence with reScaffolded: false.

How do I enable opt-in ai-badger skills in a project?▼

Add a top-level include key in .ai-badger/config.json with the skills array, then re-run the scaffold. The group name documentation expands to three skills; listing all eight opt-in names explicitly is the clearest form.

Why did my Hermes skills disappear after scaffolding in a worktree?▼

Hermes skill links are symlinks into the scaffold target directory. Scaffolding in a task worktree repoints them there, and task_tracker finish deletes the worktree, leaving dangling links. Re-run the scaffold against the durable checkout or call relink_hermes_skills to repair.

Why does the pre-commit scaffold-freshness-guard fail with unstaged changes?▼

The pre-commit hook stashes unstaged changes before running, so the guard compares against the old tree and reports false drift on exactly the unstaged files. Stage all modified scaffold outputs together in one commit to avoid the failure.

What should I exclude from an ai-badger refresh commit?▼

Exclude .ai-badger/state.json carrying other tasks' tracker state, transient .bak-<timestamp> backups, untracked review docs from parallel sessions, and the .ai-badger.bckp safety directory. Stage only the managed refresh footprint with an explicit path list.

Can I rely on the second refresh run's clean report?▼

No. The first invocation performs the re-scaffold; the second reports reScaffolded: false with zero drift even though git status shows changes. Parse the full JSON from the first run and verify with git diff rather than trusting the confirmation run.

Related Skills

Explore More Software Engineering Skills→
algoliaalgolia
9

algobot-cli

Create and manage Algolia AI agents with the Algobot CLI.

Official
Advanced
heliohqheliohq
88

handoff

Coordinates GitHub pull requests from commit to ready-for-merge state.

Official
Advanced
jokerman89jokerman89

usage-log

Records skill invocations to a JSONL audit log and generates usage frequency reports.

Community
Intermediate
fxfx

upstream-contrib

Moves a local change from a consumer repository into its upstream dependency and rewires the consumer onto the released version.

Community
Advanced