Skip to content

hgxszhj/https-github.com-zhaoxuya520-reverse-skill / skills

reverse-skill-router

Flagged

Routes reverse engineering, exploitation, penetration testing, malware, mobile, firmware, browser automation, documentation, and security tasks to the appropriate specialist skill. Use when a task spans modules or the correct reverse-skill entrypoint is unclear.

FileKindSize
CONTRIBUTING.mdmarkdown20.4 KB
INDEX.mdmarkdown11.2 KB
MASTER-ROUTING.mdmarkdown7.3 KB
SKILL.mdskill md12.9 KB
config/routing.jsonconfig14.7 KB
field-journal/2026-05-15-cellular-pro-mumu-ksad-fragment-fix.mdmarkdown1.3 KB
field-journal/2026-05-15_lumine-go-reverse.mdmarkdown2.2 KB
field-journal/2026-05-16_pentest-personalblog-fun-mass-assignment.mdmarkdown5.1 KB
field-journal/2026-05-17_pentest-vue-spa-actuator-leak.mdmarkdown3.0 KB
field-journal/2026-05-25_pentest-cf-access-sibling-subdomain-cookie-poisoning.mdmarkdown12.0 KB
field-journal/2026-05-26_pentest-newapi-rate-limit-bypass.mdmarkdown2.9 KB
field-journal/2026-06-29_burp-mcp-full-test-and-fix.mdmarkdown7.4 KB
field-journal/2026-07-05_dsl-vm-captcha-reverse.mdmarkdown3.5 KB
field-journal/2026-07-14_android-arm64-self-extract-source-recovery.mdmarkdown5.7 KB
field-journal/2026-07-18_gin-juice-client-friction.mdmarkdown1.6 KB
field-journal/2026-07-20_reverse-toolchain-windows-bootstrap.mdmarkdown3.9 KB
field-journal/2026-07-22_electron-bytenode-privileged-update-chain.mdmarkdown6.7 KB
field-journal/2026-08-01_nextjs-cdk-contract-faithful-local-reconstruction.mdmarkdown7.6 KB
field-journal/2026-08-01_pentest-encryption-oracle-public-template-sql-admin-takeover.mdmarkdown8.1 KB
field-journal/2026-08-06_cortex-m-msc-firmware-self-keyed-rotate-xor.mdmarkdown4.8 KB
field-journal/2026-08-08_client-neutral-structured-routing-pr-integration.mdmarkdown4.1 KB
field-journal/2026-08-08_pr-value-triage-and-safe-integration.mdmarkdown1.7 KB
field-journal/2026-08-14_windows-powershell-native-exit-code-pr-review.mdmarkdown5.5 KB
field-journal/2026-08-17_tool-index-r2-fallback-powershell-pwsh-host-fix.mdmarkdown12.1 KB
field-journal/2026-08-20_apk-flutter-banner-ad-removal.mdmarkdown6.0 KB
field-journal/2026-09-03_issue-pr-security-boundary-integration.mdmarkdown2.9 KB
field-journal/CONTRIBUTE-BACK.mdmarkdown5.3 KB
field-journal/_index.mdmarkdown6.1 KB
field-journal/_template.mdmarkdown4.4 KB
field-journal/anonymization.mdmarkdown6.6 KB
field-journal/precedent-auth.mdmarkdown1.7 KB
field-journal/precedent-pentest.mdmarkdown25.0 KB
field-journal/precedent-reverse.mdmarkdown13.2 KB
field-journal/seed-001_elf-packed-loader.mdmarkdown3.3 KB
field-journal/seed-002_go-malware-stripped.mdmarkdown3.4 KB
field-journal/seed-003_web-api-auth-bypass.mdmarkdown3.2 KB
field-journal/seed-004_js-sign-webpack.mdmarkdown3.6 KB
field-journal/seed-005_ad-certipy-esc1.mdmarkdown2.0 KB
field-journal/seed-006_ssrf-cloud-metadata.mdmarkdown2.7 KB
field-journal/seed-007_ntlm-relay-coercer.mdmarkdown2.9 KB
field-journal/seed-008_apk-okhttp-ssl-pin-bypass.mdmarkdown4.4 KB
field-journal/seed-009_ios-jailbreak-detect-bypass.mdmarkdown4.6 KB
field-journal/seed-010_ctf-pwn-rop-x64.mdmarkdown5.0 KB
field-journal/seed-011_pcap-protocol-reverse.mdmarkdown4.5 KB
field-journal/seed-012_log4shell-jndi-rce.mdmarkdown5.1 KB
field-journal/seed-013_kerberoasting-spn.mdmarkdown4.6 KB
field-journal/seed-014_unity-il2cpp-reverse.mdmarkdown5.3 KB
field-journal/seed-015_iot-firmware-uart.mdmarkdown6.0 KB
field-journal/seed-016_k8s-container-escape.mdmarkdown6.6 KB
field-journal/seed-017_xxe-oob-exfil.mdmarkdown5.3 KB
ops/IDENTITY.mdmarkdown2.7 KB
ops/README.mdmarkdown1.3 KB
ops/analysis-blindspot-cookbook.mdmarkdown4.1 KB
ops/analysis-decision-framework.mdmarkdown5.4 KB
ops/evidence-finding-path.mdmarkdown4.3 KB
ops/role-map.mdmarkdown3.1 KB
ops/sandbox-profile.mdmarkdown2.1 KB
ops/scope-contract.mdmarkdown3.9 KB
ops/skill-supply-chain.mdmarkdown3.6 KB
ops/timeline-workitem.mdmarkdown3.5 KB
references/community-security-skills.mdmarkdown7.5 KB
references/domain-coverage-map.mdmarkdown2.9 KB
routing.mdmarkdown27.1 KB
routing_zh.mdmarkdown29.5 KB
scripts/append-evidence.ps1script7.1 KB
scripts/bootstrap-manifest.jsonconfig12.5 KB
scripts/bootstrap-reverse.ps1script41.0 KB
scripts/bootstrap-reverse.shscript34.1 KB
scripts/case-guard.ps1script3.9 KB
scripts/case-guard.shexecutablescript3.5 KB
scripts/case-init.ps1script14.4 KB
scripts/case-init.shexecutablescript12.3 KB
scripts/consolidate-evidence.ps1script2.7 KB
scripts/consolidate_evidence.pyscript3.1 KB
scripts/extract-summaries.ps1script5.5 KB
scripts/lib/BootstrapSupplyChain.ps1script7.1 KB
scripts/lib/HostRuntime.ps1script1.9 KB
scripts/lib/RouteScope.ps1script718 B
scripts/lib/ToolDiscovery.ps1script44.6 KB
scripts/lib/WorkRoot.ps1script876 B
scripts/master-route.ps1script7.8 KB
scripts/master-route.shexecutablescript5.4 KB
scripts/refresh-tool-index.ps1script7.7 KB
scripts/refresh-tool-index.shscript17.2 KB
scripts/scan-leaks.ps1script2.9 KB
scripts/smoke.ps1script8.9 KB
scripts/test-bash-workflow.shexecutablescript2.7 KB
scripts/test-bootstrap-codex-encoding.ps1script5.0 KB
scripts/test-bootstrap-manifest.shscript12.5 KB
scripts/test-bootstrap-supply-chain.ps1script10.9 KB
scripts/test-client-neutral-bootstrap.ps1script3.5 KB
scripts/test-client-neutral-bootstrap.shscript9.5 KB
scripts/test-consolidate-evidence.ps1script3.1 KB
scripts/test-p0-friction.ps1script22.4 KB
scripts/test-parse-contracts.ps1script7.6 KB
scripts/test-routing.ps1script3.8 KB
scripts/test-routing.shscript4.8 KB
scripts/test-workflow-title-safety.ps1script1.8 KB
scripts/update-star-history.ps1script5.4 KB
scripts/verify-doc-facts.ps1script4.2 KB
scripts/verify-doc-links.pyscript3.6 KB
scripts/verify-repository-security.pyscript7.9 KB
scripts/verify-routing-coherence.ps1script33.7 KB
tests/routing-benchmark.jsonconfig17.3 KB
tool-index.md.templateother2.8 KB

scripts/test-client-neutral-bootstrap.sh

View on GitHub
#!/usr/bin/env bash
set -euo pipefail

SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
BOOTSTRAP="$SCRIPT_DIR/bootstrap-reverse.sh"
REFRESH="$SCRIPT_DIR/refresh-tool-index.sh"
REAL_HEAD="$(command -v head)"
SCRATCH="$(mktemp -d /tmp/reverse-client-neutral-XXXXXX)"
trap 'rm -rf "$SCRATCH"' EXIT

HOME_DIR="$SCRATCH/home"
BIN_DIR="$SCRATCH/bin"
TOOLS_DIR="$SCRATCH/tools"
CLAUDE_CFG="$SCRATCH/client/claude.json"
CODEX_CFG="$SCRATCH/client/codex.toml"
mkdir -p "$HOME_DIR" "$BIN_DIR" "$TOOLS_DIR" "$(dirname "$CLAUDE_CFG")"

for name in bash date dirname mktemp python3 rm sed tr uname; do
  ln -s "$(command -v "$name")" "$BIN_DIR/$name"
done

# Keep PWNLIB_NOTERM probe-local: a script-level export must leak into this
# downstream pipeline stage and be rejected, while `PWNLIB_NOTERM=1 pwn ...`
# leaves head's environment clean.
cat > "$BIN_DIR/head" <<'STUB'
#!/bin/bash
: "${TEST_REAL_HEAD:?TEST_REAL_HEAD must point to the host head executable}"
: "${TEST_HEAD_SENTINEL_MARKER:?TEST_HEAD_SENTINEL_MARKER must be set}"
if [[ "${PWNLIB_NOTERM+x}" == "x" ]]; then
  if [[ ! -e "$TEST_HEAD_SENTINEL_MARKER" ]]; then
    printf 'head sentinel: PWNLIB_NOTERM leaked outside the pwntools probe (value=%q)\n' \
      "${PWNLIB_NOTERM-}" >&2
    : > "$TEST_HEAD_SENTINEL_MARKER"
  fi
  exit 97
fi
exec "$TEST_REAL_HEAD" "$@"
STUB
chmod +x "$BIN_DIR/head"

for name in node npm npx; do
  cat > "$BIN_DIR/$name" <<'STUB'
#!/usr/bin/env bash
if [[ "${1:-}" == "--version" ]]; then echo 1.0.0; fi
exit 0
STUB
  chmod +x "$BIN_DIR/$name"
done

cat > "$BIN_DIR/ghidra" <<'STUB'
#!/usr/bin/env bash
exit 0
STUB
chmod +x "$BIN_DIR/ghidra"

cat > "$BIN_DIR/pwn" <<'STUB'
#!/bin/bash
if [[ $# -ne 1 || "$1" != "version" ]]; then
  printf 'pwn test stub: expected exactly one argument: version; got:' >&2
  printf ' <%s>' "$@" >&2
  printf '\n' >&2
  exit 64
fi
if [[ "${PWNLIB_NOTERM:-}" != "1" ]]; then
  printf '%s\n' 'Warning: _curses.error: setupterm: could not find terminfo database' >&2
fi
printf '%s\n' '[*] Pwntools v4.15.0' >&2
STUB
chmod +x "$BIN_DIR/pwn"

EMPTY_VERSION_DIR="$SCRATCH/empty-version"
EMPTY_VERSION_REFRESH="$EMPTY_VERSION_DIR/refresh-tool-index.sh"
EMPTY_VERSION_MD="$EMPTY_VERSION_DIR/tool-index.md"
EMPTY_VERSION_JSON="$EMPTY_VERSION_DIR/tool-index.json"
mkdir -p "$EMPTY_VERSION_DIR"
cp "$REFRESH" "$EMPTY_VERSION_REFRESH"
python3 - "$REFRESH" "$EMPTY_VERSION_REFRESH" <<'PY'
import hashlib
import sys
from pathlib import Path

source_path = Path(sys.argv[1])
fixture_path = Path(sys.argv[2])
source_bytes = source_path.read_bytes()
if fixture_path.read_bytes() != source_bytes:
    raise SystemExit("empty-version fixture must begin as an exact production copy")

old = b'  "node|core-runtime|Node.js runtime for MCP bridges|node|node --version|"'
new = b'  "node|core-runtime|Node.js runtime for MCP bridges|node||"'
matches = source_bytes.count(old)
if matches != 1:
    raise SystemExit(
        "production refresh must contain exactly one canonical node row; "
        f"found {matches}"
    )

fixture_path.write_bytes(source_bytes.replace(old, new, 1))
if hashlib.sha256(source_path.read_bytes()).digest() != hashlib.sha256(source_bytes).digest():
    raise SystemExit("production refresh changed while creating empty-version fixture")
PY

export PATH="$BIN_DIR:$PATH"
export HOME="$HOME_DIR"
export REVERSE_SKILL_TOOLS_DIR="$TOOLS_DIR"
export CLAUDE_MCP_CONFIG="$CLAUDE_CFG"
export CODEX_CONFIG_PATH="$CODEX_CFG"

MD="$SCRATCH/tool-index.md"
JSON="$SCRATCH/tool-index.json"
(
  unset TERM PWNLIB_NOTERM
  PATH="$BIN_DIR" TEST_REAL_HEAD="$REAL_HEAD" \
    TEST_HEAD_SENTINEL_MARKER="$SCRATCH/baseline-head-leak" \
    "$BIN_DIR/bash" "$REFRESH" "$MD" "$JSON" >/dev/null
)
if [[ -e "$SCRATCH/baseline-head-leak" ]]; then
  echo "head sentinel marker recorded a leaked PWNLIB_NOTERM in baseline refresh" >&2
  exit 1
fi

set +e
(
  unset TERM PWNLIB_NOTERM
  PATH="$BIN_DIR" TEST_REAL_HEAD="$REAL_HEAD" \
    TEST_HEAD_SENTINEL_MARKER="$EMPTY_VERSION_DIR/head-leak" \
    "$BIN_DIR/bash" "$EMPTY_VERSION_REFRESH" \
      "$EMPTY_VERSION_MD" "$EMPTY_VERSION_JSON" >/dev/null
) 2> "$EMPTY_VERSION_DIR/refresh.stderr"
empty_version_refresh_status=$?
set -e

empty_version_validation_status=0
if [[ -e "$EMPTY_VERSION_DIR/head-leak" ]]; then
  echo "head sentinel marker recorded a leaked PWNLIB_NOTERM in empty-version refresh" >&2
  empty_version_validation_status=1
fi
if [[ $empty_version_refresh_status -ne 0 ]]; then
  printf 'empty version_spec refresh must exit 0; got %d\n' \
    "$empty_version_refresh_status" >&2
  while IFS= read -r line; do
    printf 'empty version_spec stderr: %s\n' "$line" >&2
  done < "$EMPTY_VERSION_DIR/refresh.stderr"
  empty_version_validation_status=1
else
  set +e
  python3 - "$EMPTY_VERSION_MD" "$EMPTY_VERSION_JSON" <<'PY'
import json
import sys
from pathlib import Path

markdown_lines = Path(sys.argv[1]).read_text(encoding="utf-8").splitlines()
node_rows = [
    [cell.strip() for cell in line.strip()[1:-1].split("|")]
    for line in markdown_lines
    if line.startswith("| node |")
]
if len(node_rows) != 1:
    raise SystemExit(f"expected exactly one node Markdown row, got {len(node_rows)}")
if len(node_rows[0]) != 8 or node_rows[0][5] != "—":
    raise SystemExit(
        "empty node version_spec must use the existing Markdown default '—'; "
        f"got {node_rows[0]!r}"
    )

data = json.loads(Path(sys.argv[2]).read_text(encoding="utf-8"))
node_records = [tool for tool in data.get("tools", []) if tool.get("name") == "node"]
if len(node_records) != 1:
    raise SystemExit(f"expected exactly one node JSON record, got {len(node_records)}")
node_record = node_records[0]
if "version" not in node_record:
    raise SystemExit(
        "empty node version_spec expected explicit JSON null field; "
        "version field is missing"
    )
if node_record["version"] is not None:
    raise SystemExit(
        "empty node version_spec expected explicit JSON null field; "
        f"got {node_record['version']!r}"
    )
PY
  empty_version_validation_status=$?
  set -e
fi

set +e
python3 - "$JSON" "$BIN_DIR/ghidra" "$BIN_DIR/pwn" <<'PY'
import json, os, sys
data = json.load(open(sys.argv[1], encoding='utf-8'))
tools = data['tools']
assert sum(t['name'] == 'binwalk' for t in tools) == 1, 'binwalk must appear exactly once'
by_tool = {t['name']: t for t in tools}
assert by_tool['npx']['available'] is True
assert by_tool['jshookmcp']['available'] is False, 'npx must not masquerade as jshookmcp'
assert by_tool['reqable-mcp']['available'] is False, 'npx must not masquerade as reqable-mcp'
assert by_tool['ghidra']['available'] is True, 'the distro-provided ghidra launcher must be discovered'
assert os.path.realpath(by_tool['ghidra']['path']) == os.path.realpath(sys.argv[2])
pwntools_matches = [tool for tool in tools if tool.get('name') == 'pwntools']
if len(pwntools_matches) != 1:
    raise SystemExit(
        f"expected exactly one pwntools tool record, got {len(pwntools_matches)}"
    )
pwntools = pwntools_matches[0]
expected_pwntools = {
    'available': True,
    'version': '[*] Pwntools v4.15.0',
    'source': 'command',
    'skill': 'reverse-engineering',
    'purpose': 'CTF pwn exploit development framework',
}
for field, expected in expected_pwntools.items():
    if pwntools.get(field) != expected:
        raise SystemExit(
            f"pwntools {field} must be {expected!r}, got {pwntools.get(field)!r}"
        )
if os.path.realpath(pwntools.get('path') or '') != os.path.realpath(sys.argv[3]):
    raise SystemExit(
        "pwntools path must resolve to the strict pwn stub: "
        f"got {pwntools.get('path')!r}"
    )
by_cap = {c['name']: c for c in data['capabilities']}
assert by_cap['jshookmcp']['ready'] is False
assert by_cap['reqable-mcp']['ready'] is False
PY
baseline_validation_status=$?
set -e

if [[ $empty_version_validation_status -ne 0 || $baseline_validation_status -ne 0 ]]; then
  exit 1
fi

cat > "$CODEX_CFG" <<'EOF'
[mcp_servers.jshook]
command = "npx"
args = ["-y", "@jshookmcp/jshook@0.3.4"]
EOF
(
  unset TERM PWNLIB_NOTERM
  TEST_REAL_HEAD="$REAL_HEAD" \
    TEST_HEAD_SENTINEL_MARKER="$SCRATCH/registered-head-leak" \
    bash "$REFRESH" "$MD" "$JSON" >/dev/null
)
if [[ -e "$SCRATCH/registered-head-leak" ]]; then
  echo "head sentinel marker recorded a leaked PWNLIB_NOTERM in registered refresh" >&2
  exit 1
fi
python3 - "$JSON" <<'PY'
import json, sys
data = json.load(open(sys.argv[1], encoding='utf-8'))
cap = {c['name']: c for c in data['capabilities']}['jshookmcp']
assert cap['mcp_registered'] is True, 'Codex-only MCP registration must be discovered'
assert cap['runtime_available'] is True
assert cap['ready'] is True, 'registered npm MCP + npx runtime should be ready'
PY

rm -f "$CLAUDE_CFG" "$CODEX_CFG"
default_out="$(bash "$BOOTSTRAP" jshookmcp --skip-refresh)"
[[ "$default_out" == *'"status":"registration-required"'* || "$default_out" == *'"status": "registration-required"'* ]]
[[ ! -e "$CLAUDE_CFG" ]]
[[ ! -e "$CODEX_CFG" ]]

codex_out="$(bash "$BOOTSTRAP" jshookmcp --skip-refresh --mcp-host=codex)"
[[ "$codex_out" == *'"status":"ready"'* || "$codex_out" == *'"status": "ready"'* ]]
[[ ! -e "$CLAUDE_CFG" ]]
grep -Eq '^\[mcp_servers\.jshook\]$' "$CODEX_CFG"

rm -f "$CLAUDE_CFG" "$CODEX_CFG"
claude_out="$(bash "$BOOTSTRAP" jshookmcp --skip-refresh --mcp-host=claude)"
[[ "$claude_out" == *'"status":"ready"'* || "$claude_out" == *'"status": "ready"'* ]]
[[ -f "$CLAUDE_CFG" ]]
[[ ! -e "$CODEX_CFG" ]]
python3 - "$CLAUDE_CFG" <<'PY'
import json, sys
data = json.load(open(sys.argv[1], encoding='utf-8'))
assert 'jshook' in data.get('mcpServers', {})
PY

echo 'client-neutral Bash bootstrap/discovery regression passed'
rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit.Stars: 469kUpdated Claude CodeCodexVerifiedInstall
rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure.Stars: 469kUpdated Claude CodeCodexVerifiedInstall
rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics.Stars: 469kUpdated Claude CodeCodexVerifiedInstall
ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first.Stars: 194kUpdated Claude CodeCodexVerifiedInstall