{"id":"GHSA-24p2-j2jr-386w","summary":"psd-tools: Compression module has unguarded zlib decompression, missing dimension validation, and hardening gaps","details":"## Summary\n\nA security review of the `psd_tools.compression` module (conducted against the `fix/invalid-rle-compression` branch, commits `7490ffa`–`2a006f5`) identified the following pre-existing issues. The two findings introduced and **fixed** by those commits (Cython buffer overflow, `IndexError` on lone repeat header) are excluded from this report.\n\n---\n\n## Findings\n\n### 1. Unguarded `zlib.decompress` — ZIP bomb / memory exhaustion (Medium)\n\n**Location**: `src/psd_tools/compression/__init__.py`, lines 159 and 162\n\n```python\nresult = zlib.decompress(data)          # Compression.ZIP\ndecompressed = zlib.decompress(data)    # Compression.ZIP_WITH_PREDICTION\n```\n\n`zlib.decompress` is called without a `max_length` cap. A crafted PSD file containing a ZIP-compressed channel whose compressed payload expands to gigabytes would exhaust process memory before any limit is enforced. The RLE path is not vulnerable to this because the decoder pre-allocates exactly `row_size × height` bytes; the ZIP path has no equivalent ceiling.\n\n**Impact**: Denial-of-service / OOM crash when processing untrusted PSD files.\n\n**Suggested mitigation**: Pass a reasonable `max_length` to `zlib.decompress`, derived from the expected `width * height * depth // 8` byte count already computed in `decompress()`.\n\n---\n\n### 2. No upper-bound validation on image dimensions before allocation (Low)\n\n**Location**: `src/psd_tools/compression/__init__.py`, lines 138 and 193\n\n```python\nlength = width * height * max(1, depth // 8)   # decompress()\nrow_size = max(width * depth // 8, 1)           # decode_rle()\n```\n\nNeither `width`, `height`, nor `depth` are range-checked before these values drive memory allocation. The PSD format (version 2 / PSB) permits dimensions up to 300,000 × 300,000 pixels; a 4-channel 32-bit image at that size would require ~144 TB to hold. While the OS/Python allocator will reject such a request, there is no early, explicit guard that produces a clean, user-facing error.\n\n**Impact**: Uncontrolled allocation attempt from a malformed or adversarially crafted PSB file; hard crash rather than a recoverable error.\n\n**Suggested mitigation**: Validate `width`, `height`, and `depth` against known PSD/PSB limits before entering decompression, and raise a descriptive `ValueError` early.\n\n---\n\n### 3. `assert` used as a runtime integrity check (Low)\n\n**Location**: `src/psd_tools/compression/__init__.py`, line 170\n\n```python\nassert len(result) == length, \"len=%d, expected=%d\" % (len(result), length)\n```\n\nThis assertion can be silently disabled by running the interpreter with `-O` (or `-OO`), which strips all `assert` statements. If the assertion ever becomes relevant (e.g., after future refactoring), disabling it would allow a length mismatch to propagate silently into downstream image compositing.\n\n**Impact**: Loss of an integrity guard in optimised deployments.\n\n**Suggested mitigation**: Replace with an explicit `if` + `raise ValueError(...)`.\n\n---\n\n### 4. `cdef int` indices vs. `Py_ssize_t size` type mismatch in Cython decoder (Low)\n\n**Location**: `src/psd_tools/compression/_rle.pyx`, lines 18–20\n\n```cython\ncdef int i = 0\ncdef int j = 0\ncdef int length = data.shape[0]\n```\n\nAll loop indices are C `signed int` (32-bit). The `size` parameter is `Py_ssize_t` (64-bit on modern platforms). The comparison `j \u003c size` promotes `j` to `Py_ssize_t`, but if `j` wraps due to a row size exceeding `INT_MAX` (~2.1 GB), the resulting comparison is undefined behaviour in C. In practice, row sizes are bounded by PSD/PSB dimension limits and are unreachable at this scale; however, the mismatch is a latent defect if the function is ever called directly with large synthetic inputs.\n\n**Impact**: Theoretical infinite loop or UB at \u003e2 GB row sizes; not reachable from standard PSD/PSB parsing.\n\n**Suggested mitigation**: Change `cdef int i`, `j`, `length` to `cdef Py_ssize_t`.\n\n---\n\n### 5. Silent data degradation not surfaced to callers (Informational)\n\n**Location**: `src/psd_tools/compression/__init__.py`, lines 144–157\n\nThe tolerant RLE decoder (introduced in `2a006f5`) replaces malformed channel data with zero-padded (black) pixels and emits a `logger.warning`. This is the correct trade-off over crashing, but the warning is only observable if the caller has configured a log handler. The public `PSDImage` API does not surface channel-level decode failures to the user in any other way.\n\n**Impact**: A user parsing a silently corrupt file gets a visually wrong image with no programmatic signal to check.\n\n**Suggested mitigation**: Consider exposing a per-channel decode-error flag or raising a distinct warning category that users can filter or escalate via the `warnings` module.\n\n---\n\n### 6. `encode()` zero-length return type inconsistency in Cython (Informational)\n\n**Location**: `src/psd_tools/compression/_rle.pyx`, lines 66–67\n\n```cython\nif length == 0:\n    return data   # returns a memoryview, not an explicit std::string\n```\n\nAll other return paths return an explicit `cdef string result`. This path returns `data` (a `const unsigned char[:]` memoryview) and relies on Cython's implicit coercion to `bytes`. It is functionally equivalent today but is semantically inconsistent and fragile if Cython's coercion rules change in a future version.\n\n**Impact**: Potential silent breakage in future Cython versions; not a current security issue.\n\n**Suggested mitigation**: Replace `return data` with `return result` (the already-declared empty `string`).\n\n---\n\n## Environment\n\n- Branch: `fix/invalid-rle-compression`\n- Reviewed commits: `7490ffa`, `2a006f5`\n- Python: 3.x (Cython extension compiled for CPython)","aliases":["CVE-2026-27809","PYSEC-2026-2262"],"modified":"2026-07-13T07:26:55.444033267Z","published":"2026-02-26T15:20:51Z","database_specific":{"severity":"MODERATE","github_reviewed":true,"github_reviewed_at":"2026-02-26T15:20:51Z","nvd_published_at":"2026-02-26T00:16:26Z","cwe_ids":["CWE-190","CWE-409","CWE-617","CWE-704","CWE-755","CWE-789"]},"references":[{"type":"WEB","url":"https://github.com/psd-tools/psd-tools/security/advisories/GHSA-24p2-j2jr-386w"},{"type":"ADVISORY","url":"https://nvd.nist.gov/vuln/detail/CVE-2026-27809"},{"type":"WEB","url":"https://github.com/psd-tools/psd-tools/commit/6c0a78f195b5942757886a1863793fd5946c1fb1"},{"type":"PACKAGE","url":"https://github.com/psd-tools/psd-tools"},{"type":"WEB","url":"https://github.com/psd-tools/psd-tools/releases/tag/v1.12.2"}],"affected":[{"package":{"name":"psd-tools","ecosystem":"PyPI","purl":"pkg:pypi/psd-tools"},"ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"0"},{"fixed":"1.12.2"}]}],"versions":["0.1.1","0.1.2","0.1.3","0.1.4","0.10","0.2","0.5","0.6","0.7","0.7.1","0.8","0.8.1","0.8.2","0.8.3","0.8.4","0.9","0.9.1","1.0","1.1","1.10.0","1.10.1","1.10.10","1.10.11","1.10.12","1.10.13","1.10.2","1.10.3","1.10.4","1.10.5","1.10.6","1.10.7","1.10.8","1.10.9","1.11.0","1.11.1","1.12.0","1.12.1","1.2","1.3","1.4","1.8.10","1.8.11","1.8.12","1.8.13","1.8.14","1.8.15","1.8.16","1.8.17","1.8.18","1.8.19","1.8.20","1.8.21","1.8.22","1.8.23","1.8.24","1.8.25","1.8.26","1.8.27","1.8.28","1.8.29","1.8.30","1.8.31","1.8.32","1.8.33","1.8.34","1.8.35","1.8.36","1.8.37","1.8.38","1.8.8","1.8.9","1.9.0","1.9.1","1.9.10","1.9.11","1.9.12","1.9.13","1.9.14","1.9.15","1.9.16","1.9.17","1.9.18","1.9.19","1.9.2","1.9.20","1.9.21","1.9.22","1.9.23","1.9.24","1.9.26","1.9.27","1.9.28","1.9.29","1.9.3","1.9.30","1.9.31","1.9.32","1.9.33","1.9.34","1.9.4","1.9.5","1.9.6","1.9.7","1.9.8","1.9.9"],"database_specific":{"source":"https://github.com/github/advisory-database/blob/main/advisories/github-reviewed/2026/02/GHSA-24p2-j2jr-386w/GHSA-24p2-j2jr-386w.json"}}],"schema_version":"1.9.0","severity":[{"type":"CVSS_V4","score":"CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:N/VI:H/VA:H/SC:N/SI:N/SA:N/E:U"}]}