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.
npx agentmods add skills/auerbachb/claude-code-config/admin-mergenpx skills add auerbachb/claude-code-config --skill admin-mergegit clone --depth 1 https://github.com/auerbachb/claude-code-configWhat 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 | $0.00094 | $0.03807 |
| Opus 5 | $0.00047 | $0.01903 |
| Sonnet 5 | $0.00019 | $0.00761 |
| Haiku 4.5 | $0.00009 | $0.00381 |
Grade A, and why
admin-merge 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.
How it starts
The opening of the file, as written. The whole thing — 159 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Merge a PR whose only remaining blocker is branch protection on a solo-owner repo — either by auto-running the no-protection-change shape, or by printing the protection-toggle command for the user.
Portable helper resolution
Before any step, resolve the required helper. Stop visibly if it is unavailable.
resolve_script() {
local name="$1" candidate
for candidate in \
"$HOME/.claude/skills-worktree/.claude/scripts/$name" \
"$HOME/.claude/scripts/$name" \
".claude/scripts/$name"; do
if [[ -x "$candidate" ]]; then echo "$candidate"; return 0; fi
done
return 1
}
ADMIN_MERGE_SH=$(resolve_script admin-merge.sh || true)
[[ -n "$ADMIN_MERGE_SH" ]] || { echo "ERROR: admin-merge.sh not found (checked all three paths) — admin merge unavailable" >&2; exit 1; }
The safety boundary (non-negotiable) — a split, not a blanket
Claude's safety rules permanently prohibit Claude from modifying branch protection — non-negotiable, applies to every user and repo regardless of authorization or workflow history. That prohibition is about protection modification, not about the --admin flag, and the two bypass shapes fall on opposite sides of it (issues #720, #754):
| Shape | Condition | Command | Authority |
|---|---|---|---|
| plain | enforce_admins: false + required_status_checks.strict: true + verified clean-BEHIND |
gh pr merge --squash --admin — nothing else |
Claude auto-runs it via --auto-plain |
| toggle | enforce_admins: true |
DELETE …/enforce_admins → merge → POST …/enforce_admins |
Print-only. The user runs it. |
- Claude MAY execute the plain shape via
"$ADMIN_MERGE_SH" <PR> --auto-plain. It modifies no setting and steps over a staleness requirement, not a review; every safety condition is re-verified programmatically immediately before the merge. No user turn is required. - Claude NEVER executes
gh api -X DELETE .../protection/enforce_admins,gh api -X POST .../protection/..., or any other protection-modification command — not directly, not via a subprocess, not via--execute. For the toggle shape Claude only ever prints (or clipboards) the command. - The skill invokes
admin-merge.shwith--auto-plain,--print, or--launch-terminal. It must never invoke--execute(that mode can run the toggle dance and is reserved for the user in their own terminal). --auto-plainis structurally incapable of the toggle shape: its branch contains no protection-modifying call, and any non-plain shape is a hard refusal (exit 8) that falls back to printing. Mechanism:.claude/reference/admin-merge-auto-plain.md.
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 · 159 lines · 94 tokens per session scan A 887bbcff8d8b
admin-merge is a skill published in the GitHub repository auerbachb/claude-code-config (5 stars, last pushed 3d ago), licensed MIT. It adds 94 tokens to every session and 3,807 once invoked, about $0.0005 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 2026-08-31.
Other skills, from other repositories
recon
Codebase reconnaissance agent for Bug Hunter. Maps architecture, identifies trust boundaries, classifies files by risk priority, and detects service boundaries. Does NOT find bugs — finds where bugs hide.
referee
Final arbiter for Bug Hunter. Receives Hunter findings and Skeptic challenges, independently re-reads code, and delivers authoritative verdicts with CVSS scoring and proof-of-concept generation for security findings.
bughunt
Performs an exhaustive 14-dimension bug hunt across the codebase using Draft context (architecture, tech-stack, product) for false-positive elimination. Generates a severity-ranked report with code evidence, data flow traces, and suggested fixes. Optionally writes regression tests. Use when the user asks to find bugs…
juror-review
Inspect Juror Cloud PR findings and, only after an explicit confirmation, start or rerun a hosted Juror review.
coverage
Compute code coverage for active track or module. Targets 95%+ coverage with report and justification for uncovered lines. Complements TDD workflow.
tech-debt
Identify, categorize, and prioritize technical debt across seven dimensions. Generates remediation plans with effort estimates. Offered by /draft:new-track for refactor tracks.