# The algo records whose domain is context-verification

This index as JSON: https://banes-lab.com/json/api/facets/algo/domain/context-verification

## Entries

- [Phase-Separated Execution](https://banes-lab.com/records/algo/phase-separated-execution.md): Detect the current workflow phase, bind allowed operations to that phase, reject operations outside the phase contract, and emit only the artifact valid for that phase.
- [Evidence-Gated Claim Verification](https://banes-lab.com/records/algo/evidence-gated-claim-verification.md): Extract claims, resolve each claim into observable evidence requirements, collect direct implementation evidence, classify each claim as verified, contradicted, or unverified, and report discrepancies.
- [Validation Gate](https://banes-lab.com/records/algo/validation-gate.md): After each critical stage, evaluate declared success criteria, block downstream progression when critical criteria fail, and carry warning-state forward when noncritical criteria fail.
- [File Modification Recovery](https://banes-lab.com/records/algo/file-modification-recovery.md): When a file mutation fails because state changed between read and edit, reread the current file, merge the intended delta into the current content, write the complete new version, and verify persistence.
- [Trust Anchor Declaration](https://banes-lab.com/records/algo/trust-anchor-declaration.md): Declare the minimum assumptions required for the system to verify anything, bind all verification logic to those assumptions, and disclose the verification boundary.
- [Environment Capability Verification](https://banes-lab.com/records/algo/environment-capability-verification.md): Before executing advanced behavior, probe required runtime dependencies, classify each dependency failure by severity, and degrade or block capability based on criticality.
- [Tool Calibration](https://banes-lab.com/records/algo/tool-calibration.md): Create known-good and known-bad fixtures, run the verification tool against both, detect false positives and false negatives, and mark the tool reliable only if both controls pass.
- [Behavioral Self-Test](https://banes-lab.com/records/algo/behavioral-self-test.md): Execute the system’s claimed behaviors against simple positive and negative cases, compare actual output to expected output, and treat mismatch as implementation evidence failure.
- [Adversarial Input Testing](https://banes-lab.com/records/algo/adversarial-input-testing.md): Generate malicious, malformed, ambiguous, and deceptive inputs, execute the detection logic against them, and classify whether the system resists or accepts invalid patterns.
- [Defensive String Normalization](https://banes-lab.com/records/algo/defensive-string-normalization.md): Reject null input, remove dangerous path/control patterns, normalize Unicode representation, and only pass sanitized strings to filesystem, parser, or command boundaries.
- [Safe Arithmetic Contract](https://banes-lab.com/records/algo/safe-arithmetic-contract.md): Check operands before calculation, reject division by zero, reject non-finite results, enforce bounds, and return nullable or typed failure instead of unsafe numeric state.
- [Recursion Control](https://banes-lab.com/records/algo/recursion-control.md): Increment depth on recursive entry, compare against maximum depth, reject excessive recursion, and unwind depth on completion.
- [Recursive Self-Verification](https://banes-lab.com/records/algo/recursive-self-verification.md): Load the system’s own definition, extract self-claims, search for implementation evidence of each claim, classify discrepancies, and downgrade confidence when self-description exceeds implemented behavior.
- [Advanced Tool Escalation](https://banes-lab.com/records/algo/advanced-tool-escalation.md): When direct tools cannot answer a verification question, synthesize a specialized analyzer, execute it against the target, parse its output, and integrate the result as evidence.
- [Investigation Report](https://banes-lab.com/records/algo/investigation-report.md): Collect verified findings, failed checks, warnings, discrepancies, environmental limits, adversarial results, and confidence level into a structured report without performing remediation.
- [Action Log](https://banes-lab.com/records/algo/action-log.md): Accept only documented gaps as input, apply bounded changes, version the modified artifact, verify the write, and emit an action log without discovering new scope.
- [Contract-Based Verification Kernel](https://banes-lab.com/records/algo/contract-based-verification-kernel.md): Declare assumptions, detect phase, verify environment, calibrate tools, execute phase-legal behavior, test adversarially, enforce validation gates, self-verify claims, and emit a typed artifact.
- [<Context Verification Concern>](https://banes-lab.com/records/algo/context-verification-concern.md): <Detect required contract> → <Bind allowed capability> → <Execute bounded operation> → <Validate evidence> → <Emit typed result>