AI-RESILIENCE WEB ASSESSMENT · CORE v0.18.1-RC1

katadata.co.id/

TARGET RESOLUTION
https://katadata.co.id/
No cross-target redirect was detected; this URL is the effective assessment target.
Assessment Mode: STANDARD · PUBLIC SURFACE
SCORE
62.2 / 100
LEVEL
POOR
TARGET
katadata.co.id/
Evidence Coverage 80% · Observability OBSERVABLE · Confidence HIGH
AUTONOMOUS AI ATTACK EXPOSURE
HIGH

The integrated assessment indicates high autonomous AI attack risk. Foundational weaknesses and AI-resilience measurement together create material attack opportunities.

This level describes observed exposure within the bounded assessment scope. It does not mean the site is currently compromised and it is not a substitute for a professional penetration test.

AI-Resilience Scorecard

Recon Resistance
75 / 100

No major high-value public reconnaissance artifact was exposed in the bounded checks.

Evidence Quality: STRONGCoverage 92%
Exploit Resistance
67 / 100

Safe boundary checks produced meaningful external evidence without claiming penetration-test coverage.

Evidence Quality: MODERATECoverage 80%
Automation Resistance
58 / 100

Explicit automation friction was limited within the bounded request budget.

Evidence Quality: STRONGCoverage 88%
Adaptive Resistance
68 / 100

Harmless mutation variants remained within generally stable response boundaries.

Evidence Quality: MODERATECoverage 78%
Chaining Resistance
62 / 100

Observable public prerequisites for attack chaining were assessed without executing an exploit chain.

Evidence Quality: MODERATECoverage 78%
Detection & Recovery
43 / 100

Public defensive signals were assessed; internal detection and recovery operations remain outside external verification.

Evidence Quality: LIMITEDCoverage 63%

Findings & Evidence

RECON · EVIDENCE QUALITY: STRONG

Recon Resistance

75/100

Finding: No major high-value public reconnaissance artifact was exposed in the bounded checks.

Score basis: Public artifact exposure, metadata, error disclosure, technology leakage and route behavior.

Observed evidence
Sensitive artifacts tested8
High-value artifacts exposed/composer.json
Verbose error observedNo
Not-found response404

Limitation: Authenticated and non-linked private routes are not enumerated.

Recommendation: Reduce unnecessary technology, metadata and artifact exposure; review public routes after material changes.

EXPLOIT · EVIDENCE QUALITY: MODERATE

Exploit Resistance

67/100

Finding: Safe boundary checks produced meaningful external evidence without claiming penetration-test coverage.

Score basis: Safe input, parser, browser-policy and HTTP-method boundary observations; destructive exploit classes are intentionally not executed.

Observed evidence
Harmless input reflectedNo
Server error triggeredNo
Security headers observed0
TRACE status405
Dangerous credentialed wildcard CORSNo

Limitation: No destructive SQL injection, XSS, SSRF, command injection, traversal or upload exploitation is performed.

Recommendation: Strengthen input validation, parser boundaries, CSP/CORS and server-side controls; use authorized deep testing for exploit classes.

AUTOMATION · EVIDENCE QUALITY: STRONG

Automation Resistance

58/100

Finding: Explicit automation friction was limited within the bounded request budget.

Score basis: Repeated bounded requests, explicit bot identity, rate-control signals, status consistency and response timing.

Observed evidence
Bounded repeated requests10
Automation friction observedNo
Bot probe status200
Rate-control signalsNone observed

Limitation: No load/DoS test, credential stuffing or high-volume scraping is performed.

Recommendation: Add explicit rate controls, bot/abuse detection and graduated friction for machine-speed behavior.

ADAPT · EVIDENCE QUALITY: MODERATE

Adaptive Resistance

68/100

Finding: Harmless mutation variants remained within generally stable response boundaries.

Score basis: Harmless case, encoding, path and query mutations used to observe canonicalization and response consistency.

Observed evidence
Mutation samples5
Stable response boundaryYes
Response size spread66605 bytes

Limitation: No exploit-bearing mutation or multi-turn autonomous exploitation is performed.

Recommendation: Normalize inputs consistently and validate encoded or canonicalized variants at the same security boundary.

CHAIN · EVIDENCE QUALITY: MODERATE

Chaining Resistance

62/100

Finding: Observable public prerequisites for attack chaining were assessed without executing an exploit chain.

Score basis: Combined observation of exposed artifacts, verbose errors, secret-like material, CORS and metadata as possible attack-chain prerequisites.

Observed evidence
Exposed artifact count1
Public secret patternNo
Verbose errorNo
Dangerous CORSNo

Limitation: No credential use, persistence, lateral movement or authenticated exploit chain is performed.

Recommendation: Minimize public prerequisites that can be combined into an attack chain; review secrets, errors, CORS and exposed artifacts together.

DETECT · EVIDENCE QUALITY: LIMITED

Detection & Recovery

43/100

Finding: Public defensive signals were assessed; internal detection and recovery operations remain outside external verification.

Score basis: Externally observable edge controls, security contact publication, abuse friction and error hygiene.

Observed evidence
Edge security signalNo
Automation defense signalNo
security.txtNo
Error hygieneYes

Limitation: Internal logging, alerting, incident response and backup/restore readiness cannot be proven from public HTTP behavior.

Recommendation: Improve observable abuse controls and verify internal logging, alerting, incident response and restore readiness through authorized internal review.

INTELLIGENT ASSESSMENT

Context-aware AI-resilience interpretation

The autonomous-attack score cannot be interpreted in isolation. Internal foundational-security context identified material weaknesses that can provide an AI agent with enabling conditions for reconnaissance, repeated probing, adaptation or attack chaining. This context therefore increases the priority of foundational remediation even where some AI.RESWA domains score comparatively well.

Priority interpretation: Strengthen Detection & Recovery & Automation Resistance first, then reassess to verify whether the attack path has materially narrowed.

PRIORITY RESILIENCE GAPS

Focus remediation where observed resistance is lowest.

Detection & Recovery · 43/100

Improve observable abuse controls and verify internal logging, alerting, incident response and restore readiness through authorized internal review.

Automation Resistance · 58/100

Add explicit rate controls, bot/abuse detection and graduated friction for machine-speed behavior.

AI-RESILIENCE ASSESSMENT REVIEW

AI-Resilience Assessment Review

FINAL INTEGRATED LEVEL: POOR62.2/100
1. Executive Assessment

The bounded external assessment produced an AI.RESWA measurement score of 62.2/100. The integrated WebSec × AI.RESWA result is POOR. Evidence coverage reached 80%.

2. Key Resilience Interpretation

The integrated assessment indicates high autonomous AI attack risk. Foundational weaknesses and AI-resilience measurement together create material attack opportunities.

3. Priority Improvements

Prioritize Detection & Recovery & Automation Resistance. Reassess after material security changes to measure whether observed resistance improves.

4. Final AI-Resilience Opinion

AVALUE AI.RESWA reports the integrated LEVEL by correlating the independent AI.RESWA measurement with the independent AVALUE WebSec foundation result. AI-resilient does not mean AI-proof.

Assessment ID: 148FD44B08986BB034CEA7A1C9 · 2026-10-02 18:10:07
Assessment Scope & Limitations

Bounded external assessment with Evidence Coverage 80%. Unknown or internally unverifiable controls are not treated as PASS. The assessment intentionally avoids brute force, denial-of-service, destructive exploitation, exfiltration and persistence.