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 skills add pronskiy/php-src-developer --skill bug-reproduce-testgit clone --depth 1 https://github.com/pronskiy/php-src-developerWrote 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.
[](https://agentmods.dev/skills/pronskiy/php-src-developer/bug-reproduce-test)<a href="https://agentmods.dev/skills/pronskiy/php-src-developer/bug-reproduce-test"><img src="https://agentmods.dev/badge/skills/pronskiy/php-src-developer/bug-reproduce-test/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/pronskiy/php-src-developer/bug-reproduce-test"><img src="https://agentmods.dev/badge/skills/pronskiy/php-src-developer/bug-reproduce-test.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00111 | $0.02476 |
| Opus 5.5 | $0.00044 | $0.00990 |
| Sonnet 5.5 | $0.00022 | $0.00495 |
| Haiku 4.5 | $0.00011 | $0.00248 |
Grade A, and why
bug-reproduce-test 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 1mo 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 — 190 lines — stays where its author put it; the contents beside it link to each section on GitHub.
bug-reproduce-test
Second step of the bug-fix flow. Reads the Context already in SPEC.md, builds a minimal PHP reproduction, converts it into a .phpt test under $PHP_SRC_DIR, runs it, and asserts it FAILS (TDD).
Inputs
- SPEC.md in the task directory resolved per
_shared/task-dir.md(env var →.current→ cwd) withtype: bugand Context phasedone. Refuse otherwise. $PHP_SRC_DIRwith a builtsapi/cli/php.
What it produces
repro.php— minimal standalone reproduction script in the task directory, for the user's eyeballing$PHP_SRC_DIR/<tests-path>/ghNNNNN.phpt— the new test, following the conventions of the surrounding directory- Updated SPEC.md: Reproduction phase populated, status →
doneonce post-guardrail is green
Phase owned
Reproduction.
Steps this skill adds:
repro-script— minimalrepro.phpreproduces the issue against$PHP_SRC_DIR/sapi/cli/php. Notes observed vs expected.phpt-failing—.phpttest created at the appropriate path;run-tests.phpreturns failure.
Workflow
1. Pre-guardrails
- SPEC.md exists and has
type: bug. Iftype: idea, refuse with:idea SPECs don't use bug-reproduce-test. - Context phase is
done. If not, refuse with:Run /bug-context first. - Research phase soft-check: if a Research phase exists with status
todo, print one warning line and proceed:⚠ Research phase still todo. Run /research first for prior-art (or mark it `skipped` in SPEC.md to silence this warning).done,skipped, orblocked→ no warning. Statusdone→ read Research findings; they may already pin a subsystem or hint at a prior test that captures the bug. - Reproduction phase is
todoorin-progress. Ifdone, refuse with:Reproduction phase is already done. Run /bug-fix next.Ifblocked, refuse with:Reproduction is blocked: <reason>. Resolve before retrying. $PHP_SRC_DIR/sapi/cli/phpexists. If missing, do not auto-build — output the build commands and stop:$PHP_SRC_DIR/sapi/cli/php missing. Build it first: cd "$PHP_SRC_DIR" ./buildconf ./configure --enable-debug make -j$(nproc)- target_branch matches the checked-out branch. Read
target_branchfrom SPEC.md frontmatter. If set (notnull) and not equal togit -C "$PHP_SRC_DIR" rev-parse --abbrev-ref HEAD, refuse with:
Do not auto-switch branches (CLAUDE.md: "Executing actions with care" — branch swaps trigger rebuilds the user may not want unprompted).SPEC.md target_branch is <X>; $PHP_SRC_DIR is on <Y>. Switch with: git -C "$PHP_SRC_DIR" checkout <X> cd "$PHP_SRC_DIR" && make -j$(nproc) …or update target_branch in SPEC.md if the resolution was wrong. - If
target_branch: null(resolver couldn't decide), refuse with:target_branch is null in SPEC.md. Resolve the Open Question raised by /bug-context before reproducing.
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.
- 1mo ago First seen · 190 lines · 111 tokens per session scan A 819e62d53aa8
bug-reproduce-test is a skill published in the GitHub repository pronskiy/php-src-developer (8 stars, last pushed 4mo ago), licensed MIT. It adds 111 tokens to every session and 2,476 once invoked, about $0.0004 per session on Opus 5.5. 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
systematic-debugging
4-phase root cause debugging: understand bugs before fixing.
build-test
Run the project's build / typecheck / lint / test commands and emit the build.passing + tests.passing signals devloop convergence reads.
resolve-author-fix
Find the root cause, implement a focused fix, and prove behavior with targeted tests and recordings.
behavior-contract
Bug condition/postcondition formalization as testable Behavior Contracts. Defines invariants that must be preserved across fixes.
nw-bugfix
Bug fix workflow: root cause analysis → user review → regression test + fix via TDD.
08-debug
Reproduce and fix a known bug, or find an unknown root cause by hypothesis validation. Use when the user wants to fix a bug, find why something breaks, or reopen a stuck investigation. Not for building a feature or reviewing a diff.