ctf-similar-search
Skill5 starsChanged 8 months agoMITView on GitHub
Searches for similar CTF challenge writeups using Tavily API for search and content extraction. Use when the user requests finding related challenges, when solve-challenge has failed multiple times and needs reference solutions, or when first identifying a challenge type and seeking examples.
- Reads credentials
- Installs packages
1 other project uses a version of this, 1 of them edited.
Adopted by
Most recently taken up by
- ljagiello/ctf-skillsedited, first seen Oct 25, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 1
- ljagiello/ctf-skills★ 3.7k93% same
Wrote their own version · 0
None found.
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository12 other files it gives its agents.
- AGENTS.mdAGENTS.md
- ctf-ai-mlSkill
- ctf-cryptoSkill
- ctf-forensicsSkill
- ctf-malwareSkill
- ctf-miscSkill
- ctf-osintSkill
- ctf-pwnSkill
- ctf-reverseSkill
- ctf-webSkill
- ctf-writeupSkill
- solve-challengeSkill
The file
We split this skill into the rules it gives an agent
10 rules. About 0 are common, 7 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈7
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 15
Search for similar CTF challenge writeups using Tavily API.
- Distinctive: line 115
Key principle: Use keywords that uniquely identify THIS challenge, not the general technique or common patterns.
- Distinctive: line 126
If a search result looks promising, extract its full content:
- Very specific: line 145
If the challenge is already well-categorized, use the specific ctf-* skill
- Very specific: line 146
If you need OSINT techniques, use ctf-osint
- Very specific: line 147
If you need to generate a writeup after solving, use ctf-writeup
- Distinctive: line 168
Be specific: Use competition name + challenge name when available
- Distinctive: line 170
Avoid generic terms: Don't search "SSRF CTF writeup" - too broad
- Distinctive: line 171
Review first: Check search results before extracting
- Distinctive: line 172
Extract selectively: Only extract URLs that look relevant
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
CTF Similar Challenge Search
Search for similar CTF challenge writeups using Tavily API.
Prerequisites
pip install requests tavilySet your Tavily API key (from https://tavily.com) before running:
export TAVILY_API_KEY='tvly-...'API Overview
This skill provides two interfaces:
| Interface | Purpose | Command |
| Tavily Search | Search for writeups and articles | ./scripts/search.sh search <keyword> |
| Tavily Extract | Extract content from URLs | ./scripts/search.sh extract <url1> [url2] ... |
Search Workflow
Step 1: Search with Keywords
Choose keywords based on available information:
Scenario 1: Known Competition + Challenge Name
# Format: [Competition Name] [Challenge Name] — max 3 words
./scripts/search.sh search "DASCTF URL Storage"
./scripts/search.sh search "HGAME baby web"
./scripts/search.sh search "BUUCTF pwn1"Scenario 2: Challenge Name Only (no competition info)
# Use challenge name directly — max 3 words
./scripts/search.sh search "URL Storage"
./scripts/search.sh search "baby overflow"
./scripts/search.sh search "crypto-rsa"Scenario 3: Unique Identifiers (code snippets, strings)
# Use unique, non-trivial strings from the challenge — NO word limit
./scripts/search.sh search "def encrypt_flag(key): return base64"
./scripts/search.sh search "CUSTOM_MAGIC_BYTES"
./scripts/search.sh search "ERROR: invalid padding in AES block"⚠️ Critical: 3-Word Limit for Non-Code Queries
All keyword searches that are NOT code snippets or unique identifier strings MUST NOT exceed 3 words. This constraint ensures precise, focused queries that return relevant CTF writeup results instead of broad, diluted searches.
| Search Type | Word Limit | Examples |
| Competition + Challenge name | ≤ 3 words | "DASCTF URL Storage", "HGAME baby web" |
| Challenge name only | ≤ 3 words | "URL Storage", "baby overflow" |
| Code snippets / strings | No limit | "def encrypt_flag(key): return base64" |
| Error messages / identifiers | No limit | "CUSTOM_MAGIC_BYTES", "ERROR: invalid padding in AES block" |
For competition + challenge name searches, drop filler words (year numbers, "CTF", "competition", "writeup", "solution") and keep only the 3 most distinctive words:
- ❌
"HGAME 2024 baby web challenge"(5 words) - ✅
"HGAME baby web"(3 words) - ❌
"DASCTF 2023 URL Storage writeup"(5 words) - ✅
"DASCTF URL Storage"(3 words) - ❌
"PlaidCTF 2025 zero day exploit solution"(6 words) - ✅
"PlaidCTF zero day"(3 words)
For code/string searches, use the entire unique snippet — these are exempt from the word limit:
- ✅
"def xor_encrypt(data, key):"(code — exempt) - ✅
"ERROR: invalid padding in AES block"(error string — exempt) - ✅
"CUSTOM_MAGIC_BYTES"(identifier — exempt)
Keyword Rules
| ✅ Good Keywords | ❌ Bad Keywords |
| Competition + Challenge name (≤3 words) | Single vulnerability type (SSRF, XSS) |
| Challenge name (≤3 words) | Generic terms (web, pwn, crypto) |
| Unique code snippets (no word limit) | Full sentences |
| Specific error messages (no word limit) | Articles (a, the, in) |
| Library/framework names | Generic strings (flag{}, etc.) |
| Custom function names | Tool names alone |
Why avoid these?
- "SSRF" → returns generic vulnerability explanations, not CTF writeups
- "flag{" → returns millions of unrelated results
- "URL Storage DASCTF" → returns specific CTF challenges
Why the 3-word limit?
- Forces precision: only the most distinctive keywords survive
- Filters noise: filler words like "challenge", "writeup", "solution" dilute results
- Improves API efficiency: shorter, focused queries consume fewer credits
Key principle: Use keywords that uniquely identify THIS challenge, not the general technique or common patterns. Keep non-code queries to 3 words or fewer.
Step 2: Review Search Results
The search returns:
- Answer: AI-generated summary if available
- Results: List of URLs with titles and content previews
- Credits: API usage information
Step 3: Extract Content (Optional)
If a search result looks promising, extract its full content:
# Single URL
./scripts/search.sh extract "https://example.com/writeup"
# Multiple URLs
./scripts/search.sh extract "https://url1.com" "https://url2.com"Step 4: Analyze and Pivot
Based on extracted content:
- Identify common techniques used in similar challenges
- Determine if the approach is applicable
- Pivot to category-specific skill if needed
When to Pivot
- If the challenge is already well-categorized, use the specific
ctf-*skill - If you need OSINT techniques, use
ctf-osint - If you need to generate a writeup after solving, use
ctf-writeup
Quick Start
# Scenario 1: Competition + Challenge name (≤3 words)
./scripts/search.sh search "DASCTF URL Storage"
# Scenario 2: Challenge name only (≤3 words)
./scripts/search.sh search "URL Storage"
# Scenario 3: Unique identifier from challenge (no word limit)
./scripts/search.sh search "def xor_encrypt(data, key):"
./scripts/search.sh search "ERROR: invalid padding"
# Extract content from promising results
./scripts/search.sh extract "https://ctftime.org/writeup/12345"Tips
- Be specific: Use competition name + challenge name when available
- Use unique strings: Code snippets, error messages, unique identifiers
- Avoid generic terms: Don't search "SSRF CTF writeup" - too broad
- Review first: Check search results before extracting
- Extract selectively: Only extract URLs that look relevant
- Respect the 3-word limit: For non-code queries, strip filler words and keep only the 3 most distinctive keywords
$ARGUMENTS
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
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.

