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 commands/friedbotstudio/baseline/grant-pushgit clone --depth 1 https://github.com/friedbotstudio/baselineWhat 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 | $0.00027 | $0.00667 |
| Opus 5 | $0.00014 | $0.00333 |
| Sonnet 5 | $0.00005 | $0.00133 |
| Haiku 4.5 | $0.00003 | $0.00067 |
Grade A, and why
grant-push 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.
What it actually says
Write a consent token to .claude/state/push_consent so the Git Commit Guard hook allows the next git push on a protected branch. The token is the current UNIX epoch timestamp on line 1; any optional note goes on line 2.
How this works structurally: when the user typed /grant-push, the consent_gate_grant UserPromptSubmit hook ran before this body was passed to Claude and wrote a short-lived consent marker at .claude/state/.push_consent_grant. The git_commit_guard PreToolUse hook (Write matcher) reads that marker and allows Claude to write the consent file because the marker is fresh. Claude cannot forge the marker — that's what makes the gate structural. The Bash-matcher leg of the same guard then enforces the consent token on the actual git push invocation, but only when the current branch matches project.json → git.protected_branches.
State-write discipline (binding — see .claude/CONSTITUTION.md §2 "State-write discipline"). The consent token at .claude/state/push_consent is a Tier 1 consent artifact: it SHALL be written with the Write tool only. You SHALL NOT write it via Bash (no >/>> redirect, heredoc, tee, cp, or sed -i) — destructive_cmd_guard blocks Bash writes to consent paths, and the git_commit_guard Write matcher validates the marker only on the Write tool. Use Bash solely for the precheck (git rev-parse) and the epoch (date +%s).
Steps:
- Git-repo precheck. Run
git rev-parse --is-inside-work-tree 2>/dev/null. If the exit status is non-zero, this project is not a git repository: refuse to write the consent token and tell the user "Not a git repository —/grant-pushis inapplicable. Push has no meaning outside a git repo." Stop here. - Run
date +%sto get the current epoch. - Write the consent token with the Write tool, overwriting any prior token: the epoch on line 1 (and the optional note
$ARGUMENTSon line 2 if non-empty) to.claude/state/push_consent. Do not use a Bash redirect ortee— that path is guard-blocked. - Confirm to the user: "Push consent granted at , valid for 300s (until <HH:MM:SS local>). The next
git pushon a protected branch will be allowed. Pushes on branches NOT inproject.json → git.protected_branchesdo not require this consent."
Do not run git push yourself in this command. The user asks explicitly when they want a push; this command only opens the window.
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 · 22 lines · 27 tokens per session scan A ede000e21619
grant-push is a command published in the GitHub repository friedbotstudio/baseline (11 stars, last pushed 6d ago), licensed Apache-2.0. It adds 27 tokens to every session and 667 once invoked, about $0.0001 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 2026-08-30.
Other commands, from other repositories
audit-agents-skills
Audit quality of agents, skills, and commands in a Claude Code project.
security-audit
Comprehensive security audit of your project AND Claude Code configuration. Analyzes secrets exposure, injection surfaces, dependencies, hook security, and produces a scored security posture assessment.
audit-quiz-coverage
Find quiz coverage gaps from recent guide/CHANGELOG/CC-releases changes and propose new questions.
worktree
Create a git worktree for isolated feature/fix work.
sync
Check guide/landing synchronization status.
init-docs
Fetch official Anthropic Claude Code docs and store as local baseline snapshot.