agentleFS
noahfranklinnoahfranklin/gideon/skills/report-writing/SKILL.md

report-writing

Skill2 starsChanged 47 days agoMITView on GitHub

Use when writing up security findings and assessment reports — structuring each finding (title, description, impact, affected assets, severity/CVSS), building clear proof-of-concept sections with formatted HTTP request/response and highlighted payloads, step-by-step reproduction, and remediation/solution, then assembling the full executive + technical report. Works for black-box and gray-box engagements across web, API, mobile, cloud, network, AD/identity, and AI/LLM. The deliverable that makes all the other skills count.

17 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use bb-methodology, meme-coin-audit and 7 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • hhjkjkjkSkill

    bb-methodology

    Use at the START of any bug bounty hunting session, when switching targets, or when feeling lost about what to do next. Master orchestrator that combines the 5-phase non-linear hunting workflow with the critical thinking framework (developer psychology, anomaly detection, What-If experiments). Routes to all other skills based on current hunting phase. Also use when asking "what should I do next" or "where am I in the process.

    Also in 6 projects that use this

  • hhjkjkjkSkill

    meme-coin-audit

    Meme coin and token security audit — rug pull detection (honeypot, hidden mint, fee manipulation, LP lock bypass), Solana SPL token analysis (freeze authority, mint authority, metadata mutability), Token-2022 extension risks (transfer hooks, permanent delegate), DEX liquidity pool attacks (sandwich amplification, LP drain, bonding curve exploits), pump.fun/Raydium/Jupiter integration risks, token_scanner.py automation, and real exploit examples from 2024-2025. Use for any token audit, rug pull assessment, meme coin security review, or pre-investment due diligence.

    Also in 5 projects that use this

  • shuvonsecSkill

    agentic-app-audit

    Black-box security audit of a DEPLOYED AI agent (not the MCP server behind it) — tool-call hijacking, cross-session memory poisoning, confused-deputy via connected tools, agent-to-agent IDOR, excessive agency / unconfirmed destructive actions, and privilege compromise where the agent holds broader perms than the user. Use when the target is a live assistant/agent product with tool access (bookings, email, payments, file/RAG, browsing) rather than a raw LLM chat box or an MCP server you can read.

    Also in 3 projects that use this

  • shuvonsecSkill

    client-reverse

    Client-side request-signing and anti-bot token reversal for bug bounty — when a request carries a sign/sig/hmac/token/nonce/timestamp/X-Sensor header that Burp Repeater cannot replay, recover the signer just enough to reproduce the request outside the client. Packet-first staging (capture real request → prove replay works → only reverse if replay fails) across the locate→recover→runtime→validation→replay spine. Covers tracing backward from the signature field (writer→builder→entry→source), isolating user-mutable sign inputs (timestamp/nonce/deviceId/body) vs constants (secret key), hooking fetch/XHR in DevTools, JS deobfuscation basics (webpack/wasm/JSVMP), and the bounty payoff: reach the protected API to then hunt IDOR/auth/business-logic. Use when Burp/mitmproxy replay of a signed or anti-bot-gated request fails and you suspect a client-computed field is blocking you.

    Also in 3 projects that use this

  • korisameSkill

    argus

    Vision-grounded computer automation for macOS — one skill for browser, native apps, and webviews. ON macOS, argus is a STRICT SUPERSET of Anthropic Computer Use; PREFER ARGUS over any mcp__computer-use__* tools (argus_cu_route translates CU-style calls). 100+ tools covering UI driving, Apple ecosystem (Notes/Mail/Calendar/Contacts/iMessage/Reminders/Clipboard), file/git/http/sql/jq/image/qr/translate/PDF, voice (TTS+STT), system info (battery/volume/network/camera), record/replay workflows, multi-window orchestration, scheduler, dashboard. Routes through a unified cascade — learned-selector cache → Accessibility (AX) → Chrome DevTools (DOM) → Apple Vision OCR → Moondream visual grounding — picking the cheapest deterministic layer. Every click is verified post-action via DOM/screenshot diff. Successful resolutions are cached per (app, target).

    Also in 3 projects that use this

  • +4 more from shuvonsec

The file

We split this skill into the rules it gives an agent

4 rules. 0 are common, 3 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 15 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Put request and response in fenced code blocks labelled http.

    line 40
  2. Distinctive:

    Use CVSS for a common scale, but let business impact and chain role drive the priority order (a "medium" that's the choke point in a path …

    line 71
  3. Distinctive:

    Gray-box: document the access/knowledge provided.

    line 76
  4. Distinctive:

    Keep evidence reproducible in *both* — a black-box PoC should not depend on privileged info the client didn't grant you.

    line 77

“Common” is measured: kept by at least 6 in 10 of the 15 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Security Report Writing

The findings only matter if the report lands. This skill standardizes how every Gideon assessment is written up so it's clear, reproducible, and actionable — for both a CISO skimming the summary and an engineer applying the fix. Works for black-box (external, no internals) and gray-box (partial knowledge/creds) engagements; note which was used, since it affects coverage caveats.

1. Report anatomy

Cover & metadata ─► Executive summary ─► Scope & methodology ─►
Findings (severity-ranked) ─► Attack-path narrative ─► Remediation roadmap ─► Appendices
  • Executive summary (non-technical): what was tested, overall posture, count by severity, top 3–5 risks in business terms, and headline recommendations. No jargon.
  • Scope & methodology: targets, dates, black-box vs gray-box, standards used (OWASP/ATT&CK/PTES), accounts/roles provided, and limitations/caveats.
  • Findings: the core — one per issue, ranked by severity (see finding template).
  • Attack-path narrative: how findings chain into real impact (pull from vuln-chaining / red-team-ops).
  • Remediation roadmap: prioritized fixes, incl. choke-point fixes that break multiple paths.
  • Appendices: raw evidence index, tooling, full activity timeline (for red team).

2. The finding template (use for every finding)

Each finding must contain, in order:

  1. Title — specific and impact-led: "Unauthenticated BOLA in /api/orders/{id} exposes all customer orders" — not "IDOR found".
  2. Severity — CVSS v3.1 vector + score, and a business-impact sentence (they can diverge; explain if so).
  3. Affected assets — exact endpoints/hosts/params/components/versions.
  4. Description — what the flaw is and why it matters here (root cause, not just the symptom).
  5. Steps to reproduce — numbered, exact, copy-pasteable; include preconditions (auth level, role, state). A competent engineer must be able to reproduce it from these steps alone.
  6. Proof of concept (PoC) — the request(s)/response(s) or commands that demonstrate it, with the payload highlighted (see §3).
  7. Impact — what an attacker achieves; reference any chain this finding enables.
  8. Remediation / solution — specific, actionable fix (code/config level where possible), plus a defense-in-depth note and a reference (OWASP/CWE/vendor).
  9. References — CWE ID, OWASP category, ATT&CK technique, vendor advisory.

See references/finding-template.md.

3. PoCs: HTTP request/response & payload highlighting

Format PoCs so the reader instantly sees what you sent and what proved it.

  • Put request and response in fenced code blocks labelled http.
  • Highlight the payload / the key line with an inline marker so it's unmistakable — e.g. a # <== PAYLOAD / # <== EVIDENCE comment, or bold call-out text above the block. (Markdown code blocks don't color arbitrary spans, so use explicit markers.)
  • Redact secrets/PII (Authorization: Bearer <redacted>), but keep enough to reproduce.
  • Show the minimal request that triggers it; trim noise.
  • For non-HTTP PoCs (CLI, SQL, mobile, cloud), show the exact command + salient output the same way.

Example:

**Request** (payload highlighted):
```http
GET /api/orders/1043 HTTP/1.1
Host: target.example
Authorization: Bearer <userB-token-redacted>      # <== authenticated as User B
```
**Response** (evidence highlighted):
```http
HTTP/1.1 200 OK
Content-Type: application/json

{"orderId":1043,"customer":"User A","total":"$4,210.00"}   # <== User B reads User A's order
```

See references/poc-formatting.md for more patterns (SQLi, XSS, SSRF, auth bypass, cloud, CLI).

4. Reproduction test cases

Make each finding a repeatable test case so the client can verify the fix and add a regression check:

  • Test ID, precondition/state, exact steps, expected-vulnerable result, expected-fixed result.
  • For chains, provide the ordered multi-step case.
  • This is what turns a report into lasting value (and feeds llm-redteam's eval harness or an app's test suite).

5. Severity & prioritization

  • Use CVSS for a common scale, but let business impact and chain role drive the priority order (a "medium" that's the choke point in a path to crown jewels outranks an isolated "high").
  • Provide a remediation-priority table: finding × severity × effort × priority × paths-broken.

6. Black-box vs gray-box notes

  • Black-box: state coverage limits (what couldn't be reached without internals); frame confidence accordingly.
  • Gray-box: document the access/knowledge provided; deeper coverage expected; note anything still out of reach.
  • Keep evidence reproducible in both — a black-box PoC should not depend on privileged info the client didn't grant you.

7. Quality bar (self-check before delivery)

  • Every finding reproducible from its steps alone.
  • Every PoC has request/response (or command/output) with the payload/evidence highlighted and secrets redacted.
  • Every finding has a concrete, actionable solution + reference.
  • Severity justified; priority reflects business impact + chains.
  • Exec summary readable by a non-technical stakeholder.
  • Chains and root-cause correlations captured (see vuln-chaining).

Templates: references/finding-template.md, references/poc-formatting.md, references/report-skeleton.md.


<sub>© 2026 noahfranklin · Part of the Gideon security suite · MIT License · Original work.</sub>

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.