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/boxlogodev/sapstack/sap-session-startgit clone --depth 1 https://github.com/BoxLogoDev/sapstackWhat 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 | $0.00040 | $0.01857 |
| Opus 5 | $0.00020 | $0.00928 |
| Sonnet 5 | $0.00008 | $0.00371 |
| Haiku 4.5 | $0.00004 | $0.00186 |
Grade A, and why
sap-session-start 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 — 161 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/sap-session-start — 새 Evidence Loop 세션 시작
입력: $ARGUMENTS
이 커맨드는 Evidence Loop의 Turn 1 INTAKE를 수행합니다. 라이브 SAP 접근 없이 운영자가 가져온 증거를 기반으로 진단을 시작하는 엔트리 포인트입니다.
🧭 동작 단계
1. 인자 파싱
- 포지셔널 인자: 증상 설명 (최대 2000자). 자유 텍스트.
- 옵션 플래그:
--country {iso}— ISO 3166-1 alpha-2 (kr, de, us, jp, ...).config.yaml에 기본값 있으면 생략 가능.--release {code}— ECC6_EhP7, S4_2022 등. 스키마의 enum 참조.--deployment {on_premise|private_cloud|public_cloud}--language {ko|en|...}— 세션 전체 언어--intake-files <path>...— 초기 Evidence Bundle 파일들 (선택적)--reporter-role {end_user|operator|consultant}
2. 세션 ID 생성
sess-{YYYYMMDD}-{6자리 영숫자}
예: sess-20260412-m2p9xt
3. 파일 시스템 준비
.sapstack/sessions/{session_id}/
├── state.yaml ← 새로 생성
├── bundles/ ← 빈 디렉토리
├── hypotheses/ ← 빈 디렉토리
├── requests/ ← 빈 디렉토리
├── verdicts/ ← 빈 디렉토리
└── files/ ← 빈 디렉토리
4. state.yaml 초기화
schemas/session-state.schema.yaml에 따라 다음 필드를 채워 새 파일 작성:
session_id,schema_version: "1.0.0",created_at,last_updated_atcreated_by.role(기본값: operator)originating_surface: clistatus: intakeinitial_symptom:description: 파싱된 포지셔널 인자reporter_role: 플래그 or 기본값language: 플래그 or 기본값 (config.yaml → ko)country_iso: 플래그 or config.yaml 기본값matched_symptom_index_entry:data/symptom-index.yaml과 fuzzy 매칭된 ID (매칭 없으면 null)
sap_context: 플래그와 config.yaml 병합turns:- 첫 엔트리:
turn_number: 1,turn_type: intake,status: active,surface: cli
- 첫 엔트리:
current_turn_number: 1audit_trail:- 첫 엔트리:
session_created
- 첫 엔트리:
5. 초기 Bundle 처리 (선택적)
--intake-files가 제공되면 즉시 Bundle 1개를 생성해 bundles/evb-*.yaml로
저장하고, 각 파일을 files/ 아래로 복사하며 SHA-256 해시를 계산해
evidence-bundle.schema.yaml에 맞는 메타데이터를 작성합니다.
자동 감지:
.txt,.log→kind: transaction_log또는message_text.csv,.xlsx→kind: table_export.png,.jpg→kind: screenshot.html→kind: fiori_page_html
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 · 161 lines · 40 tokens per session scan A 05132cf5d987
sap-session-start is a command published in the GitHub repository BoxLogoDev/sapstack (18 stars, last pushed 2d ago), licensed MIT. It adds 40 tokens to every session and 1,857 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 commands, from other repositories
eval-author
Draft a new eval case (eval/cases/ .json) + scaffold its golden folder, goldenpending until captured on the VM.
eval-improve
Work the eval-loop corpus — rank failure clusters, fix the top actionable defect, validate against held-out, open a PR.
eval-run
Run an eval case end-to-end on the VM (implement → build → score → record → roll back). VM/full-mode only.
research-l2
Executes Phases 1-3 of the L2 process of the abapwiki knowledge base on a slice: launches the abap-functional-researcher sub-agent (gap discovery + multi-source auto-research: wiki -> raw/docs -> MCP abap-fs on the system read-only -> standard knowledge), ingests gaps and evidence, and generates questionnaires for…
ingest-l0
L0 bootstrap of the abapwiki knowledge base: creates stubs for all custom objects from the TADIR. Use this skill to initialize the repo from scratch or to ingest a new TADIR export. Deterministic operations (no LLM sub-agent), idempotent.
ingest-l1
L1 analysis loop for the abapwiki knowledge base: for each batch it launches the abap-analyzer sub-agent in parallel, then the adversarial judge abap-deepcheck (separate session), applies only the analyses that pass the fail-closed gate, and commits. Resumes exactly after an interruption. Use this skill to document…