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/azrtydxb/procoder/prgit clone --depth 1 https://github.com/azrtydxb/procoderWhat 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 | $0.00020 | $0.01084 |
| Opus 5 | $0.00010 | $0.00542 |
| Sonnet 5 | $0.00004 | $0.00217 |
| Haiku 4.5 | $0.00002 | $0.00108 |
Grade A, and why
pr 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 — 81 lines — stays where its author put it; the contents beside it link to each section on GitHub.
The user invoked /procoder:pr with arguments: $ARGUMENTS
Every procoder command below is the procoder binary on PATH.
- Read .procoder/github/WORKFLOW.md and follow it — the repo's rules win
over this skill's defaults. If it is missing, get the default via
procoder templates, write it, then follow it. Default rule: feature work happens in a git worktree (one per branch/agent), never dirtying the default-branch checkout — if this work isn't in one yet, note it and keep the branch clean.
0b. If this branch closes a tracked issue, check whether a pull request is
already open for it: gh pr list --state open --search "<issue number>",
and the issue's own timeline. Compare each result's head branch against
this one. The only PR that is not a reason to stop is the one whose head
IS this branch — push to that rather than opening a second one. Anything
else, including your own older branch and an agent working under the same
account, is somebody already on it: review theirs instead. A second PR
for the same fix wastes the reviewer and, when the other one is a
contributor's, costs the contributor.
- Run
procoder gitandprocoder check. Fix everything BLOCKING before going further — a PR that fails its own gate wastes the reviewer. The gate'sinfoimpact lines are the change's blast radius: open each referencing file it names and verify the change breaks nothing there (or runprocoder index impactfor the full list). - Read the real diff (
git diff <default-branch>...HEAD) and the commits. Summarise what actually changed, not what you remember intending.
2a. The docs-impact question — MANDATORY, answered in writing before the
self-review: "What does this change alter about what a reader must be
told?" Features change the story, not just the reference. Answer with
either the list of pages updated (README narrative, docs site pages, the
command reference) or the explicit sentence "no reader-visible change
because ". A PR without this answer does not get opened — docs
lag was our longest-lived escaped-finding class, and mention-checks alone
did not catch it (the gate's README must mention families are the
mechanical floor, this question is the judgment ceiling).
2b. The pre-PR self-review — the first fresh pair of eyes is OURS, not the downstream reviewer's:
- Read .procoder/github/REVIEW.md (missing →
procoder templatesprints the default; write it first). - Dispatch a FRESH-context reviewer subagent — not yourself; the author's context hides the author's blind spots — with: the rubric verbatim, the branch diff, and the instruction to report findings as file:line, what breaks, and the fix, ending with a severity-counted verdict or exactly "Nothing found — open the PR."
- Give that same agent the SECOND lens in the same pass, quoted from
/procoder:simplify the way the rubric is quoted: the five tags with
their definitions, the one-line finding format
<file>:L<line>: <tag> <what>. <replacement>., the rule against inventing a finding, and the exact null resultLean already. Ship.Scope it to the DIFF — the repo sweep belongs to /procoder:release. - Report the simplify findings first, the correctness verdict last.
- Fix every Critical/Important finding (commit them), decide each cut —
taking it or saying why not, P-CONTROL — re-run
procoder check, and only then continue. Downstream bot reviews are the fallback net — anything they catch later becomes a lesson (see /procoder:merge's reflection step).
- Fill .procoder/github/PULL_REQUEST_TEMPLATE.md section by section from that
diff. If the template is missing, get it via
procoder templates, write it, then fill it. - Write the drafted title and body to a temp file and run
procoder scrub <file>. It must say clean — no Co-Authored-By, no "generated with", nothing presenting the work as an AI's. Fix and re-scrub until clean. - Keep the title at 72 characters or fewer — it becomes the squash commit's subject and our own gate flags longer ones.
- Show the user the final title and body BEFORE creating anything.
- Create the PR with
gh pr create --title ... --body-file .... If gh is missing, runprocoder initfirst. - Hand off to /procoder:merge for the gates-and-reviews phase, or offer to run it now.
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 · 81 lines · 20 tokens per session scan A d61afcdedf49
pr is a command published in the GitHub repository azrtydxb/procoder (196 stars, last pushed 2d ago), licensed Apache-2.0. It adds 20 tokens to every session and 1,084 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
setup
Initialize a new project with SDLC-compliant structure. Creates required files, configures build system, sets up CI/CD, and establishes quality tooling.
validate
Run SDLC compliance check against the current project. Validates build system, code quality, testing, CI/CD, security, documentation, VCS, and release configurations.
run-tests
Run automated functional tests using the hook-driven test framework. Execute the test suite to validate all project functionality.
test-report
Generate test execution report in markdown format.
add-test
Interactively add a new test definition to the test suite.
release-swarm
Orchestrate complex software releases using AI swarms that handle everything from changelog generation to multi-platform deployment.