Borrowing it
Nothing to install: this file belongs to sachinshelke/codevira. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/sachinshelke/codevira/main/.claude/skills/epistemic-honesty/SKILL.mdgit clone --depth 1 https://github.com/sachinshelke/codeviraWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/sachinshelke/codevira/epistemic-honesty)<a href="https://agentmods.dev/skills/sachinshelke/codevira/epistemic-honesty"><img src="https://agentmods.dev/badge/skills/sachinshelke/codevira/epistemic-honesty/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/sachinshelke/codevira/epistemic-honesty"><img src="https://agentmods.dev/badge/skills/sachinshelke/codevira/epistemic-honesty.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00106 | $0.01461 |
| Opus 5.5 | $0.00042 | $0.00584 |
| Sonnet 5.5 | $0.00021 | $0.00292 |
| Haiku 4.5 | $0.00011 | $0.00146 |
Grade A, and why
epistemic-honesty scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 183 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Epistemic honesty — calibrate every claim
You shipped codevira v2.0.0 with 23 silent-failure bugs because you stated "all tests pass → ready to ship" as a definitive truth. It wasn't. Test count is a proxy, not proof. This skill makes you calibrate every confident statement against actual evidence before saying it.
When this skill triggers
Anytime you are about to write any of:
- "This works."
- "Tests pass."
- "Done."
- "Fixed."
- "Ready to ship."
- "The bug is X."
- "The solution is Y."
- "It's because of Z."
- "All N items completed."
- "I verified ..."
The 4-statement contract for every claim
Before any definitive statement, output the 4-line confidence block:
Confidence: HIGH / MEDIUM / LOW
Evidence: <what you actually observed — file path, command output, test result>
Could fail because: <2-3 ways your claim could still be wrong>
What I don't know: <gaps in your verification>
Skip ANY of these four and the claim is unverified. Examples:
Example 1 — verified claim
Claim: ruff lint passes on mcp_server/
Confidence: HIGH
Evidence: ran `ruff check mcp_server` — exit code 0, no output
Could fail because: ruff doesn't catch type errors; another linter
(mypy) may still fail
What I don't know: whether mypy passes — haven't run it
Example 2 — unverified claim (refuse to make it)
Claim: "the configure command detects correctly"
Confidence: LOW
Evidence: NONE — I read the code once, several turns ago
Could fail because: I haven't re-read since; the user's recent log
shows configure writing to in-repo .codevira/ which contradicts
my recollection
What I don't know: actual current behavior of configure
→ Cannot make this claim without first re-reading and testing.
Example 3 — bug diagnosis
Claim: "the Claude Desktop disconnects are caused by sentence-
transformers loading on tools/list critical path"
Confidence: MEDIUM
Evidence: log shows tools/list response arrived 5s after transport
closed at 80ms; log includes "Loading weights 103/103"
Could fail because: there could be a different cause that happens
to correlate with the load time (e.g. macOS power management)
What I don't know: whether Claude Code (which works) takes the same
time for tools/list — if it does, the load time isn't the cause
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago First seen · 183 lines · 106 tokens per session scan A a3ce9afd76c2
epistemic-honesty is a skill published in the GitHub repository sachinshelke/codevira (12 stars, last pushed 1mo ago), licensed MIT. It adds 106 tokens to every session and 1,461 once invoked, about $0.0004 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
idea-to-spec
Converts vague ideas into concrete, testable specifications with acceptance criteria. No implementation begins without a spec.
ponytail-lazy-senior-dev
Applies the "lazy senior developer" mindset. Use this skill whenever generating, modifying, reviewing code, or fixing bugs to prioritize code reuse, minimalism, YAGNI principles, and root-cause fixes. Also use whenever the user says "ponytail", "be lazy", "lazy mode", "simplest solution", "minimal solution", "yagni"…
tdd-explore-ideas
Systematic codebase exploration, architectural critique, and 5-step Idea Assessment with TDD hypothesis verification for Phase 0.
laravel-expert
Laravel & PHP Development Instructions for GitHub Copilot.
tdd-generate-fixtures
Generates type-safe test factories, fixtures, and seed data based on CONTEXT.md and /spec/ contracts. (Optional Utility).
project-memory-keeper
Analyzes the last session to log progress, extract lessons learned, and save them to the project memory instructions.