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/ahepi/deepreason/dr-execute-stepnpx skills add AHepi/DeepReason --skill dr-execute-stepgit clone --depth 1 https://github.com/AHepi/DeepReasonWhat 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 | $0.00048 | $0.01887 |
| Opus 5 | $0.00024 | $0.00944 |
| Sonnet 5 | $0.00010 | $0.00377 |
| Haiku 4.5 | $0.00005 | $0.00189 |
Grade A, and why
dr-execute-step 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 — 143 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Execute one step
Input: CHECKLIST.md. Output: one more checked step with its done-criterion output pasted beneath it. You do this for ONE step, then return. The loop lives in the orchestrator, not in you — that is what keeps a long change from drifting.
Procedure
- Re-read REQUEST.md (including Amendments) and CHECKLIST.md in full. Find the FIRST unchecked step. That is your entire job. Do not read ahead "to be efficient"; do not batch steps.
- Confirm the step still makes sense against the tree (a prior step may have failed silently). If the tree contradicts the step — file missing, test already passing, root identity occupied — do not improvise: record the contradiction under the step, commit, and return to the orchestrator (route: dr-plan-steps).
- Execute the action. Only files this step's spec item names may change. Mid-step discoveries ("this file also needs...") go to PARKED.md or, if the change cannot land without them, back through dr-spec-change as an amendment — never just typed in.
- Run the done-criterion command. Paste its real output (trimmed to
the relevant lines) under the step. If it does not match expected:
the step is NOT done — leave it unchecked, record the output and
one line on the mismatch, and return to the orchestrator. Two
failures of the same step = stop condition, in the standard format
— canonical in
dr-drive-harness§6's calibration note. - If this step changed behaviour, update the map in the SAME
commit — see "Map obligations" below. If it changed the packaging
surface (pyproject entry points, CLI commands, MCP tools/schema,
wheel layout), update
scripts/wheel_smoke.py's pinned expectations and re-run the smoke in the same commit too — no gate runs it for you. - Mark the box, update CHECKLIST.md — including its header State:
line (next step, blockers), which is what a fresh session resumes
from — and if the step is tagged [COMMIT] (or changed any file):
git addthis step's files, then runpython tools/diff_budget.py <tranche-base> --ceiling <SPEC.md's ceiling> --paths <SPEC.md's declared areas>and read itsDIFF_BUDGET_RESULT_V1.verdict. WITHIN/NO_CEILING: continue. EXCEEDED is a STOP in the standard format (decision, priced options, recommendation), not a footnote — an estimate-only ceiling trips on nothing unless read. Alongside it, runpython tools/blast_radius.py --files <this step's actually git-added files> --symbols <this step's actually touched top-level defs, from the diff hunks> --against <tranche-base>(Rung G6,docs/map/INV-frozen-surfaces.md) and diff itsfrozen_surface_contacts/reachabilityoutput against THIS document's own Frozen-surface contact forecast and Blast-radius census sections in SPEC.md. Anyfrozen_surface_contactsentry not already named in SPEC.md, or anyreachabilityentry whosedirectionisnewly_dead/newly_liveand was not predicted, is DRIFT — a STOP in the exact same format asdiff_budget.py's own EXCEEDED, never a footnote (docs/ERRATA_EXECUTOR.md, "the frozen-surface stop did not hold"). No drift: continue. Then commit and push 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 · 143 lines · 48 tokens per session scan A 19d7e854f445
dr-execute-step is a skill published in the GitHub repository AHepi/DeepReason (140 stars, last pushed 2d ago), licensed MIT. It adds 48 tokens to every session and 1,887 once invoked, about $0.0002 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 skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
brainstorming
You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.
chat-pet-sprite-creation
Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…
agent-host-chat-contributions
Build and review cross-cutting agent-host chat behavior through lifecycle contributions. Use when adding turn lifecycle side effects, prompt or context injection, restored-history transformation, protocol-action observation, or when reviewing changes that add code to AgentSideEffects or AgentService.
auto-perf-optimize
Run agent-driven VS Code performance or memory investigations. Use when asked to launch Code OSS, automate a VS Code scenario, run the Chat memory smoke runner, capture renderer heap snapshots, take workflow screenshots, compare run summaries, or drive a repeatable scenario before heap-snapshot analysis.