jrag: Skill for Claude Code

.claude/skills/publish-pip/SKILL.md

publish-pip is a skill for Claude Code from HumanBean17/jrag. It costs 73 tokens per session (3,503 once invoked), scanned C, original, MIT.

A release runbook for publishing the jrag-cli and java-codebase-rag Python packages to PyPI, the main public package index for Python.

In plain words
What is it for?
Use it to prepare a tag-triggered CI release, publish both packages together, or follow the manual fallback when CI is unavailable or the PyPI state needs reconciliation.
Why use it?
It keeps both package names at the same version and protects against PyPI's permanent version rule, where an uploaded version cannot be replaced.

Skill for Claude Code

Written for Claude Code: disable-model-invocation in frontmatter.

This is HumanBean17/jrag's own configuration. It tells Claude Code how to work on jrag itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything jrag configures →

Needs its repository: it reads a path above its own folder, which exists only inside the repository. The line is ( cd shim && rm -rf dist build && ../.venv/bin/python -m build ).

Reuse

Borrowing it

Nothing to install: this file belongs to HumanBean17/jrag. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/HumanBean17/jrag/master/.claude/skills/publish-pip/SKILL.md
Clone the repo
git clone --depth 1 https://github.com/HumanBean17/jrag

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for publish-pip

README.md
[![agentmods](https://agentmods.dev/badge/skills/humanbean17/jrag/publish-pip.svg)](https://agentmods.dev/skills/humanbean17/jrag/publish-pip)
Your own site
<a href="https://agentmods.dev/skills/humanbean17/jrag/publish-pip"><img src="https://agentmods.dev/badge/skills/humanbean17/jrag/publish-pip.svg" alt="Measured on agentmods" height="20"></a>
Per session 73 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 3,503 The whole file, excluding the scripts and references it only reads on demand.
Security scan C 2 findings. A grade says what 26 rules found in the file — not that it is safe.
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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00073 $0.03503
Opus 5 $0.00036 $0.01751
Sonnet 5 $0.00015 $0.00701
Haiku 4.5 $0.00007 $0.00350

Measured 7d ago against content hash 6eff34c43fd0, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-07, from the pricing page.

Security

Grade C, and why

publish-pip scanned grade C with 2 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 7d 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.

Recursive force deletehighDestructive command

rm -rf with a variable or a broad path is one typo away from removing the wrong tree.

**not** a bare glob: under zsh (default `NOMATCH`), `rm -rf dist build *.egg-info`

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

6. **Verify both names on PyPI** via the JSON API. ⚠️ Python's `urllib`/`requests`
.claude/skills/publish-pip/SKILL.md · 239 lines

How it starts

The opening of the file, as written. The whole thing — 239 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Publish Pip Package

Dual-name PyPI release runbook for this repo. Every release ships under BOTH PyPI namesjrag-cli (canonical) and java-codebase-rag (legacy shim) — at the same version; never leave them diverged. PyPI uploads are permanent: a version can be yanked but never overwritten, so the version is verified before any upload.

Two paths, in priority order:

  1. Primary — tag → CI publishes both names. The maintainer runs scripts/release.sh; the release.yml workflow builds, guards, and dual-publishes via OIDC Trusted Publishing, verifies both names, and opens the GitHub Release. This is the normal path.
  2. Manual fallback — when CI is down or reconciling a diverged PyPI state. Build + upload both dists from a worktree with the venv build / twine tools. The legacy name comes from the shim/pyproject.toml rebuild, NOT a sed name-swap (that procedure is obsolete).

Primary release path — tag → CI publishes both names

The maintainer runs the human-half orchestrator; CI does the rest.

.venv/bin/bash scripts/release.sh X.Y.Z

release.sh bumps root + shim/pyproject.toml in lockstep (via scripts/bump_version.py --apply), runs the install-artifact sync gate, commits bump version to X.Y.Z, lays an annotated tag vX.Y.Z, and pushes both to origin (--atomic, so the commit and tag land together or not at all). It builds and uploads nothing itself. --dry-run prints the plan without mutating; --no-push commits+tags without pushing; --skip-sync-check bypasses the artifact-sync gate (temp-repo only).

The CI half.github/workflows/release.yml, fires on v* tags — then:

  1. builds the canonical jrag-cli dist (python -m build),
  2. guards it with scripts/check_dist_version.py --dist dist --pyproject pyproject.toml,
  3. publishes jrag-cli to PyPI via OIDC Trusted Publishing (no stored token),
  4. builds the shim inside shim/ (python -m build with working-directory: shim),
  5. guards it with check_dist_version.py --dist shim/dist --pyproject shim/pyproject.toml,
  6. publishes java-codebase-rag (order is fixed: AFTER the canonical dist is live, so the shim's jrag-cli==<ver> dep pin resolves),
  7. verifies BOTH names report the tag's version (fails the job if either is missing), and
  8. opens a GitHub Release with auto-categorized notes (.github/release.yml).

Read the full file on GitHub · 239 lines

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.

  1. 7d ago First seen · 239 lines · 73 tokens per session scan C 6eff34c43fd0

Subscribe to this mod's changes

publish-pip is a skill published in the GitHub repository HumanBean17/jrag (9 stars, last pushed yesterday), licensed MIT. It adds 73 tokens to every session and 3,503 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it C with 2 findings (recursive force delete, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

document

Run /document pr | changelog | release-note | postmortem (or let it ask) to write the human facing prose about a change. Drafts from the real commits and diff, writing to the right place. Does not write code, tests, or specs.

jsmastery-pro/skills · 65 tokens

solon-development-skill

Solon Java framework expert (NOT Spring). Use for Solon apps, Solon AI (ChatModel/RAG/MCP/Agent/Harness/Talent), Solon Flow, Solon Cloud, Nami RPC, SqlUtils/MyBatis, and Solon annotations (@Mapping, @Inject, @SolonMain, @Component). Independent IoC/AOP and plugins — never use Spring annotations or spring-boot…

opensolon/soloncode · 91 tokens

dxkit-onboard

Walk a customer through setting up dxkit on a repo from scratch — checks state, installs, scaffolds, configures hooks, runs doctor, fixes any gaps, captures the first baseline, sets up branch protection + Codespaces prebuild. Also drives a DETERMINISTIC deep-configuration pass — it runs vyuh-dxkit configure --plan…

vyuh-labs/dxkit · 245 tokens

dxkit-pr

Open a pull request with a title + body grounded in the branch's real commits and diff — what changed, features implemented, findings fixed — plus a reviewer checklist and the dxkit guardrail/allowlist/score signals a reviewer needs. Use when the user says "raise a PR", "open a pull request", "create the PR", "write…

vyuh-labs/dxkit · 99 tokens

dxkit-schema

Configure, read, and act on the dxkit model-schema drift gate — list the data-model inventory, preview drift before pushing, explain why the guardrail blocked a model change, and ship a deliberate breaking change the safe way (migration + expiring accepted-risk allowlist entry). Use when the user says "set up the…

vyuh-labs/dxkit · 116 tokens

dxkit-uninstall

Cleanly and non-intrusively remove dxkit from a repo, restoring its exact pre-dxkit state — reverse every additive merge (settings.json, CLAUDE.md, .gitignore, package.json), delete every file dxkit created, and clean up hooks + CI + the .dxkit/ tree. Dry-run first. Use when the user says "remove dxkit", "uninstall…

vyuh-labs/dxkit · 123 tokens