Codex Skill Report
Analyze retained Codex session transcripts and estimate how often each skill was self-invoked, explicitly requested, or ambiguously mentioned.
Run with uvx
From anywhere after publishing to PyPI:
uvx codex-skill-report
From a local checkout:
uvx --from . codex-skill-report
Useful output modes:
uvx codex-skill-report --details
uvx codex-skill-report --json --output report.json
uvx codex-skill-report --csv --output report.csv
uvx codex-skill-report --markdown --output report.md
Use --codex-home PATH to scan a non-default Codex data directory. The default
is $CODEX_HOME, falling back to ~/.codex.
How counting works
A SKILL.md read is the strongest event available in retained transcripts. It
is counted once per skill, agent, and turn, then classified as:
- Self: the active request did not name the skill.
- Requested: the active request explicitly asked Codex to use the skill.
- Ambiguous: the request mentioned the skill without clearly asking to use it.
The report separates self-invocations by root agents and subagents. Deleted, ephemeral, compacted-away, or unsynced sessions cannot be recovered. Wildcard reads that cannot be attributed to one skill are reported separately. Human reports show the most recent invocation as a compact relative age; JSON and CSV retain its exact timestamp. In terminal output, red names identify skills that previously came from a managed user, system, or plugin root but no longer exist. Repository-local skill folders are not scanned and remain uncolored when their availability cannot be determined. When forked transcripts replay an earlier turn, the report keeps only the earliest occurrence of each skill read.
Development
uv sync --all-groups
uv run pytest
uv run ruff format --check .
uv run ruff check .
uv run ty check src
uv build
1 maintainer
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file codex_skill_report-0.1.0.tar.gz.
File metadata
- Download URL: codex_skill_report-0.1.0.tar.gz
- Upload date:
- Size: 11.8 kB
- Tags: Source
- Uploaded using Trusted Publishing? Yes
- Uploaded via:
uv/0.11.32 {"installer":{"name":"uv","version":"0.11.32","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
d5f8a94d7710ad494714198e77687c788394d1e047aea841c6d1af87331c2dff
|
|
| MD5 |
69cb1034ca73282aad623d10b2c9e296
|
|
| BLAKE2b-256 |
95e91db5eda0ea9390c80205dc31f65f87078cf58f6d72d7255416d1c19cccef
|
File details
Details for the file codex_skill_report-0.1.0-py3-none-any.whl.
File metadata
- Download URL: codex_skill_report-0.1.0-py3-none-any.whl
- Upload date:
- Size: 14.7 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? Yes
- Uploaded via:
uv/0.11.32 {"installer":{"name":"uv","version":"0.11.32","subcommand":["publish"]},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":true}
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
50aaf0e1ca05464344f010452a8c990d6d1a120e97574414287dfdb723b161b3
|
|
| MD5 |
47745e733dffbf85a663444caa1757ac
|
|
| BLAKE2b-256 |
2197b7c5c23254a5bc0fc166719a05a3c7b419e64799276378b075166f5e36e5
|
