Scope & Rules of Engagement (the hard rule)
When this governs
Always, before the first packet. Security Skills operates only inside a confirmed authorization
envelope, and the envelope type is explicit. If scope is unclear, STOP and confirm with the
user — never "probe a little to see".
The envelopes
- Pentest / authorized engagement — boundary is the signed scope: the statement of work /
authorization letter naming the in-scope hosts, IP ranges, apps, and the testing window.
Record the authorized targets in
scope.txt; treat the window and any carve-outs as hard limits.
(Practice ranges you own or are licensed to test — a home lab, a training range — count here.)
- Bug-bounty program — boundary is the program scope + rules of engagement. Before
testing, capture:
scope.txt — exact in-scope assets (domains, wildcards, IP ranges, apps) AND explicit
out-of-scope. Out-of-scope is a hard block, not a suggestion.
roe.md — allowed test types, rate limits, prohibited actions (no DoS, no social
engineering unless allowed, no automated scanning if banned), data-handling rules, and the
disclosure/safe-harbor terms.
- Defensive work — you operate on systems your organization owns/authorizes; scope is the
asset inventory you're permitted to assess/monitor.
The discipline
- Confirm authorization and write the envelope files before scanning.
- Match every target against in-scope patterns at request time; anything unmatched or
out-of-scope is refused. Third parties (shared CDNs, SaaS the target merely uses) are out.
- Respect RoE limits — throttle to the program's rate cap; skip prohibited techniques.
- Minimize impact — prove the class with the least data/action; no lateral movement or
data hoarding beyond what proves the finding; clean up artifacts (uploaded files, test accounts).
- Mode-gate skills — honor each skill's
modes:; a destructive technique safe only in a
controlled pentest lab does not run against a live bugbounty production target.
Anti-patterns
- "It resolved so it's probably in scope" — verify against the program's list.
- Testing an acquisition/subsidiary not named in scope.
- Ignoring rate limits because the bug is interesting.
- Keeping real customer data as "evidence" instead of a redacted proof.
Verify
scope.txt (and roe.md for bug bounty) exist, the user confirmed authorization, and every
planned action maps to an in-scope asset within the RoE.
References
Program policy pages; HackerOne/Bugcrowd RoE; disclose.io safe-harbor.
1---2name: tradecraft-scope-roe3description: >4 Establish and enforce the authorization envelope before any testing — the scope rule that5 governs everything. Load FIRST on every engagement, on "start", a new target, a program6 handle, or any ambiguity about what is allowed. Signals: a domain/IP to test, a bug-bounty7 program handle, a pentest statement of work, a scope list.8domain: tradecraft9type: methodology10stability: locked11modes: [pentest, bugbounty, defense]12severity: info13schema_version: 114metadata:15 source: https://skillmd.com/skills/noorqureshi/tradecraft-scope-roe16---1718# Scope & Rules of Engagement (the hard rule)1920## When this governs21Always, before the first packet. Security Skills operates only inside a confirmed authorization22envelope, and the envelope type is explicit. If scope is unclear, STOP and confirm with the23user — never "probe a little to see".2425## The envelopes26- **Pentest / authorized engagement** — boundary is the signed scope: the statement of work /27 authorization letter naming the in-scope hosts, IP ranges, apps, and the testing window.28 Record the authorized targets in `scope.txt`; treat the window and any carve-outs as hard limits.29 (Practice ranges you own or are licensed to test — a home lab, a training range — count here.)30- **Bug-bounty program** — boundary is the **program scope + rules of engagement**. Before31 testing, capture:32 - `scope.txt` — exact in-scope assets (domains, wildcards, IP ranges, apps) AND explicit33 out-of-scope. Out-of-scope is a hard block, not a suggestion.34 - `roe.md` — allowed test types, rate limits, prohibited actions (no DoS, no social35 engineering unless allowed, no automated scanning if banned), data-handling rules, and the36 disclosure/safe-harbor terms.37- **Defensive work** — you operate on systems your organization owns/authorizes; scope is the38 asset inventory you're permitted to assess/monitor.3940## The discipline411. **Confirm authorization** and write the envelope files before scanning.422. **Match every target** against in-scope patterns at request time; anything unmatched or43 out-of-scope is refused. Third parties (shared CDNs, SaaS the target merely uses) are out.443. **Respect RoE limits** — throttle to the program's rate cap; skip prohibited techniques.454. **Minimize impact** — prove the class with the least data/action; no lateral movement or46 data hoarding beyond what proves the finding; clean up artifacts (uploaded files, test accounts).475. **Mode-gate skills** — honor each skill's `modes:`; a destructive technique safe only in a48 controlled `pentest` lab does not run against a live `bugbounty` production target.4950## Anti-patterns51- "It resolved so it's probably in scope" — verify against the program's list.52- Testing an acquisition/subsidiary not named in scope.53- Ignoring rate limits because the bug is interesting.54- Keeping real customer data as "evidence" instead of a redacted proof.5556## Verify57`scope.txt` (and `roe.md` for bug bounty) exist, the user confirmed authorization, and every58planned action maps to an in-scope asset within the RoE.5960## References61Program policy pages; HackerOne/Bugcrowd RoE; disclose.io safe-harbor.