Implement fixes and correction in Fable5 Report - Part 1

[gadieid]

I have reviewed the Fable 5 code review report, and almost all the flagged items require our attention. Going forward, we will be running these code scans periodically, so it is important we establish a clean baseline and a clear process now.

Please proceed with the following priorities for this phase:

1. Security Issues (Chapter 1)

Items 1.1 – 1.3: While these are effectively false positives and not actual security threats, they will continue to trip automated code scans. This will cause policy compliance blockers down the line, so they must be refactored or mitigated to satisfy the scanners.

Items 1.4 – 1.6: These are genuine vulnerabilities and are highly important to fix in this phase.

2. Unit Testing (Chapter 5)

The issues raised in Chapter 5 tie directly into our existing goal of improving test coverage. Please link these specific fixes to Issue #943.

3. General Issues (Other Chapters)

For all remaining chapters in the report, focus exclusively on fixing items marked as Critical and Major during this current phase.

Tracking and Documentation Requirements
To maintain a clear picture of our progress, we need strict traceability for this report.

Referencing: Every commit or PR must explicitly reference the specific issue number from the scan report.

Status Mapping: We need a documented outcome for every item in the scan. Categorize each item into one of three statuses:

Fixed

Not an Issue / False Positive (Include a brief explanation of why)

Postponed (Added to the backlog for a future phase)

[ilayfalach]

Implementation complete — PR #955

We addressed 32 of the 51 Fable5 scan findings on branch issue953. All items that could be fixed without a major architectural redesign were implemented and committed. Full item-by-item status is documented in SCAN_STATUS.md in the repo.

What was done (32 fixed):

All 7 security items — credentials removed from git, passwords masked before logging, bootstrap scripts parameterized via env vars, all os.system(f-string) shell injection vectors replaced, eval() on data-controlled strings removed, sys.path validated before inserting DB-supplied paths, pickle annotated
7 data layer items — mutable default arguments fixed across 6 files, hardcoded absolute developer paths removed, getAllDocuments silent query bug fixed, getCacheDcouments typo fixed, inline angle math replaced with hera.utils helpers, raw EPSG integers replaced with WSG84/ITM constants
6 architecture items — OF_LSM registry path corrected, hidden DB write removed from getter, **kwargs added to abstractToolkit, circular import removed
3 code quality items — bare except: blocks fixed
3 + 1 testing items — MongoDB probe optimized to 1-second timeout, compare_outputs crash-swallowing removed; item 5.1 confirmed already fixed in #884
6 packaging items — README rewritten, setup.py version and dependencies fixed, TEST_UI.md paths fixed, 31 MB conda tarballs untracked, meta.yaml updated
What was NOT done (2 Critical items postponed):

Item 2.1 — import hera performs network I/O, filesystem writes, and DB writes at module import time. This requires a full lazy-connection lifecycle redesign — the DB connection cannot simply be deferred without breaking all existing consumer code. Scope is too large for this PR; tracked in a dedicated follow-up issue.

Item 5.2 — Zero test coverage for simulations/ (~70 modules) and riskassessment/ (~17 modules). Writing adequate coverage for these subsystems is a substantial independent effort, already tracked under #943. This is also in the follow-up issue.